YT YouTube2Text

YouTube2Text → Transcripts → Why the AI’s honeymoon is ending (and tech workers are feeling it) | Noam Segal

Why the AI’s honeymoon is ending (and tech workers are feeling it) | Noam Segal — Transcript

by Lenny's Podcast · 15,535 words · 2,281 segments · language en · Watch on YouTube

Open in transcriber Summarize with AI Download SRT/VTT

Full transcript

0:00Bad news for the tech community.
0:03>> Burnout at work is a major problem and
0:06AI could add to this already overwhelmed
0:09employee problem.
0:10>> The honeymoon period with AI is over.
0:13When people are asked, what what are you
0:14afraid of? Losing my job to AI is
0:17actually second to last. What we saw
0:19rise up to the top is the expectation to
0:22do more for the same pay. We did this
0:26survey on how people are feeling in tech
0:30right now. Burnout is [music] increasing
0:32significantly. Optimism is declining.
0:35It's never been crazier. When are we
0:37going to get to the optimistic part of
0:39this [laughter] uh of this episode? Half
0:42of the people in tech are feeling
0:44incredible. And the other half told us,
0:47"My brain is rotting. My work feels
0:51worse." Elon had this really interesting
0:53way of describing it. We're definitely
0:55living in a simulation if we're alive
0:57right now when we're about to start
0:58building data centers in space and AI is
1:01going to be as smart as human. What does
1:02this tell us we should be doing if we
1:03can do anything? We're in the second
1:05inning of a massive shift. No one knows
1:09how it will end, but all you can do is
1:13Today, my guest is Noam Seagull. Noam
1:16and I go a long ways back. We worked
1:18together at Airbnb for many years where
1:20he was my research partner. Since then,
1:22he went on to be a research leader at
1:24Intercom, Twitter, Wealthfront, Meta,
1:27Zapier, and Figma. Gnome is so amazing,
1:30and over the past couple years, I've
1:31been lucky to partner with him on a
1:33bunch of research projects. The most
1:35ambitious of which has been a tech
1:37worker sentiment survey, which we plan
1:40to run every year. This is the second
1:42time we've run it, and the results are
1:44so interesting. I believe this is the
1:46largest survey of its kind looking at
1:48how people in tech feel about their jobs
1:51and AI [music] and layoffs and burnout
1:54and the future of their careers. There's
1:56so much in this [music] and basically
1:58this episode is a conversation about the
2:00results. At the end of the conversation,
2:02we go through a bunch of advice on what
2:04you [music] might do if you are
2:06currently on the struggle bus and trying
2:08to figure out what to do. I am so
2:09excited for this [music] episode. I'm so
2:11excited for this report. A big thank you
2:12to Noam for doing the work to create the
2:15survey, run the survey, analyze the
2:16survey, and write the post that we will
2:18link to. It is so good. Before we get
2:20into it, do not forget to check out
2:22lennisproass.com
2:24for a free year of the hottest and most
2:26well-crafted [music] AI products in the
2:28world, available exclusively to Lenny's
2:30newsletter subscribers. With that, I
2:32bring you Noam Seagull.
2:37Noam, thank you so much for being here
2:39and welcome to the podcast. Thank you so
2:41much, Lenny. It's such an incredible
2:43experience to be here. Thanks for having
2:45me.
2:45>> Uh, it's my pleasure. We've actually
2:47worked together for a long time. And no,
2:50few people know this. We, you were the
2:51researcher on our team. Uh, when I was
2:53at Airbnb, this was like 10 years ago.
2:55>> Decades ago, a decade ago, Lenny, so
2:57much has changed.
2:58>> So much has changed.
3:00>> So much has changed. Uh, which is a
3:02great segue to what we're going to be
3:03talking about today, which is we did
3:06this survey on how people are feeling in
3:10tech right now. Uh, it's never been
3:12crazier. And so we did this report on
3:15just how are people feeling. And this is
3:17the second time we've done this. We ran
3:18this a year ago. And the idea here is
3:20we're going to have this yearly tech
3:22sentiment survey that's going to track
3:23how tech people are feeling. Before we
3:25get into the takeaways, what's kind of
3:27the give us a little context just on
3:29this survey like how how you ran it, how
3:32what were we looking at? Anything people
3:33need to know?
3:34>> Yeah, so to your point, this is the
3:35second time we're running this survey,
3:38which I really feel is unique in the
3:41tech industry. I appreciate you for
3:44giving me the opportunity and giving us
3:46an opportunity as a community to learn
3:49more about how we're all feeling in
3:51tech. I'm not aware of any survey like
3:53this at this scale looking at these
3:55things. Last year when we ran the
3:58similar tech sentiment, the first
4:00inaugural tech sentiment survey, the
4:03title was burnt out but optimistic. We
4:06were seeing fairly high levels of
4:08burnout but also fairly good levels of
4:10optimism about where things are going.
4:12And we actually followed up on the
4:14burnout point and we created the armor
4:17framework if you remember. And uh I'm
4:20I'm really glad we we did that. This
4:22year we surveyed about 6,000 people
4:27from product, engineering, design,
4:31research, pretty much uh any role,
4:32marketing, pretty much any role in tech.
4:35And the results were very interesting.
4:38So I'm very excited to jump into them.
4:40>> Yeah. I just remember when you shared
4:41the first draft of the analysis uh and I
4:44was just like reading it, I was like,
4:45"Holy this is so interesting." and
4:48it just feels like it's putting words to
4:50the way I feel, the way that I hear a
4:53lot of people feel. So, I'm really
4:54excited that we're doing this. This
4:55episode is brought to you by our
4:57season's presenting sponsor, Work OS.
5:00What do OpenAI, Anthropic, Cursor,
5:02Versel, Replet, Sierra, Clay, and
5:05hundreds of other winning companies all
5:07have in common? They are all powered by
5:09work OS. If you're building a product
5:11for the enterprise, you've felt the pain
5:13of integrating single signon, skim,
5:15arbback, audit logs, and other features
5:18required by large companies. Work OS
5:20turns those deal blockers into drop-in
5:22APIs with a modern developer platform
5:25built specifically for B2B SAS.
5:27Literally, every startup that I'm an
5:29investor in that starts to expand up
5:31market ends up working with Work OS. And
5:33that's because they are the best.
5:35Whether you are a seedstage startup
5:37trying to land your first enterprise
5:38customer or a unicorn expanding
5:40globally, work OS is the fastest path to
5:42becoming enterprise ready and unblocking
5:45growth. It's essentially Stripe for
5:47enterprise features. Visit works.com to
5:50get started or just hit up their Slack
5:51where they have actual engineers waiting
5:53to answer your questions. Workos allows
5:56you to build faster with delightful
5:57APIs, comprehensive docs, and a smooth
6:00developer experience. Go to works.com to
6:03make your app enterprise ready. Today
6:05we're going to go through the top 10
6:06takeaways. What's kind of the like the
6:09big picture takeaway from the report
6:10overall? If someone could take one big
6:12takeaway away. Well, first I'll say that
6:15you and I both noticed an incredible
6:17post from Elena Vera a couple of days
6:20before this pod was recorded and she's
6:23talking about the AI confidence theater
6:27and how in conversations these days in
6:30tech everything is dead. First of all,
6:32engine coding is dead, design is dead,
6:34SAS is dead, etc. Obviously, that's not
6:37true. So, what you're about to see in
6:40these results is the best picture we
6:43could provide and probably the best
6:45picture anyone's provided so far of how
6:48people are really feeling about AI. And
6:51to me, this also ties to the fact that
6:54if you were if you're watching this and
6:56you work in tech, you probably answer
6:58some sort of bianual culture amp style
7:01survey about how you're feeling. But the
7:04questions don't really dive any deeper
7:06than the effectiveness or of your
7:08manager or leadership at the company you
7:10work at. Whereas the big picture in this
7:15survey and and I think the the thing
7:18that stood out to us the most is how
7:22bifocated the workforce is right now.
7:25And what you're going to see is that AI
7:28has an outlier level outsiz impact on
7:34how people feel about their work more so
7:36than any other uh characteristic of
7:40their job whether it's their role, the
7:43company they work for, the company size,
7:46their level etc. Half of the people in
7:50tech are feeling incredible, energized,
7:54amplified, excited about the technology
7:58and the future of, you know, their role.
8:01And the other half feel like the future
8:05is unclear. They feel destabilized. They
8:08feel diminished. Um, other negative
8:11emotions around that. It's uh it's a
8:16very interesting picture that really
8:18cuts the tech workforce in half and
8:21that's what we'll see in a second.
8:22>> It feels so like yes, this feels correct
8:25as we we look at people in tech. Like it
8:27just uh is so clear that this happening
8:29and what's interesting is it's 50/50 in
8:32the results. I didn't expect to be so
8:33even. That was a really interesting
8:34takeaway.
8:36>> It's 50/50. When you dive a little bit
8:38deeper into the data, there is some
8:40nuance there. So, we're looking at this
8:42slide with the question that you and I
8:45both decided to answer the survey, which
8:47is, how has AI shifted your professional
8:50identity, if at all? And I'll actually
8:53call out first of all that only 3% of
8:55people reported that AI hasn't shifted
8:58their identity. So, clearly AI has
9:01changed how we each view ourselves.
9:04That's first of all. But then to your
9:06point, the most prominent thing here is
9:07that 50% of people in tech are feeling
9:11amplified. They feel like they can do
9:13more. They feel like they can do better.
9:17It's a wholly positive uh emotion.
9:22And it's great. It's great. It's great
9:24for them. But the other half of tech
9:27divide into three distinct groups. The
9:29first group is people who feel like
9:30their role is being redefined. They
9:33don't, that group doesn't actually have
9:36that much clarity or specific positive
9:40or negative emotions around what's
9:41happening. They just don't really
9:43understand it. It's a very mixed bag
9:46type of emotion where it's very clear to
9:48me that my role is fundamentally
9:51changing, but I don't really know what
9:54it means. And that's 27% of our sample.
9:59Then as we go down the line, we have two
10:03groups who have clear negative emotions
10:06around what's going on. The first group,
10:0814% of people feel destabilized.
10:11You know, the the ground is shaking
10:14beneath them. They have very little
10:16clarity on what's going on. They have
10:18high anxiety
10:20levels. They report having uh very
10:23negative and and uh pessimistic views of
10:27what's to come. And then 5% of people
10:30feel diminished.
10:32They feel like the great powers of AI
10:34have taken something away from them. Uh
10:39and something that obviously is never
10:41going to come back because what's the
10:43saying? Uh AI and air models are the
10:47worst today that they'll ever be.
10:49They're only getting better. So people
10:51who who feel diminished now will only
10:54feel more diminished in the future
10:56probably. Super. So, following that
10:59thread,
10:59>> sewing for the bad news. Sewing for the
11:01the bad news.
11:01>> There's there's some good news here. So,
11:03following the thread, uh this these
11:06buckets correlate super strongly with
11:10basically every other measure of how
11:12people are doing. So, this kind of
11:13regression on just like every other core
11:16uh question you asked and and it's so
11:17interesting to see how connected these
11:19are.
11:20>> Yeah. And just to put a couple of
11:22statistical terms on this to to clarify
11:24how big the effect is in tech we are
11:28somewhat obsessed with the notion of
11:29statistical significance. But the thing
11:31about statistical significance is that
11:33if you have a large enough sample you
11:35can sort of reach that type of
11:38significance regardless of how much the
11:42effect matters practically. A better way
11:45to look at things is using an effect
11:47size measure like Coen's D in in this
11:50case. And what effect size tells you is
11:53practically how much does this thing
11:55matter. So for example last year a
11:58couple of the largest effects we saw
12:01were number one your manager. Who your
12:04manager is and how effective they are
12:06matters a ton to how much you're
12:08enjoying your job and how optimistic you
12:10are and how burnt out you are or not.
12:12And we'll see some of that later on. And
12:15the second large effect we found was the
12:17founder happiness effect. Founders, at
12:20least people who are currently founders
12:22and in their journey of running a
12:24startup, are incredibly happy probably
12:26because they have a lot of agency and
12:29and autonomy relatively to um the
12:31typical employee.
12:33this finding, this insight around the
12:37impact of AI on your identity and on
12:41every single other variable around your
12:44job is about three times as large as
12:48those other effects.
12:51So this technology, this era that we're
12:53in is having an outlier level, outsized
12:57impact on how people are feeling about
12:59work more so than anything else we've
13:01seen. So let's actually look at that
13:03data of just how this sort of
13:05classification connects to just
13:06optimism, burnout, things like that. Uh
13:09because this is and we're going to come
13:10back to just like what exactly is that
13:12divide? Just a question to ask that
13:14tells you which side you're on. But
13:15let's look at this first. Yeah. So we'll
13:17dive into some of these things in a
13:18little more depth later on. But just to
13:20put it all together, what you're looking
13:22at here is the AI identity stance and
13:27then a breakdown by your AI identity
13:29around career optimism, burnout, layoff,
13:33worry, and an interesting one for me and
13:37we'll talk about this further. Would you
13:39recommend your role to someone coming
13:42into the industry now? And what you can
13:45see is that as you move from the people
13:48who feel amplified to the people who
13:50feel diminished, career optimism goes
13:54down significantly.
13:56Burnout or reports of burnout go up
14:00significantly.
14:01Layoff worry also rises significantly.
14:06And perhaps the most significant finding
14:08of them all is that people would not
14:13recommend
14:14junior people, early career people
14:17entering into the tech workforce at this
14:20moment given what's happening with AI.
14:24And you're seeing these very linear,
14:28clear, and organized effects, which just
14:32demonstrates to you the impact that AI
14:34is having on all of these different
14:36variables. And what I love is you took
14:39all of this kind of all these data
14:40points and you've created these kind of
14:42four archetypes of people today. And uh
14:46I feel like everyone listening is going
14:47to be like, okay, that's me. So let's go
14:50through that.
14:50>> Yeah. So, we're seeing these these four
14:53tech workers. And the way we got to this
14:55is later on in the survey, we asked
14:58people about how they're feeling and to
15:01report specific emotions. And people
15:03could choose as many emotions as they
15:06wanted, right? And by the way, fun fact,
15:09on average, people selected five
15:11different emotions. And I think a couple
15:14of people selected 13 emotions that
15:17they're feeling, which is a lot. So the
15:21first archetype we found is the
15:24energized tech worker and that's 41% of
15:29our samples. These are the people who
15:31were saying that product has become fun
15:34again. I'm exploring things. I feel like
15:38I'm in the tech amusement park where
15:42everything is open to me and I can
15:44experiment with new approaches. I'm a
15:47builder now. I have powers that I've
15:49never had before, etc. These people are
15:51very energized about the technology and
15:54about their their careers. The second
15:56group we called the conflicted, the
15:59ambivalent middle. These are 35% of the
16:03sample. These are people who are having
16:06on the one hand the most fun they've
16:08ever had as builders, as PMs, as
16:13designers, as engineers, whatever their
16:15role is. And on the other hand, they're
16:18feeling the most uncertainty that
16:20they've ever felt in their careers. They
16:23just don't understand where this is
16:26going and whether they're building the
16:30things that will ultimately lead to the
16:33end of their careers as they know it.
16:37The third archetype are the disoriented
16:40people. The people who feel like their
16:42role keeps shifting. There's a quote
16:45that we saw from one of the the
16:46respondents. We're like farmers on the
16:49cusp of the industrial revolution and we
16:52just don't see a clear path to what's
16:55happening. Very interesting comparison
16:57to to that era. And these are people who
17:01are feeling very disoriented around
17:03what's going on in their career and
17:05where things are are leading. And then
17:07finally, 12% of the sample we
17:11categorized as resentful. They're
17:13feeling pressured. They're feeling
17:16checked out. They are not enjoying their
17:19work right now. They're saying things
17:21like, "I've been forced to use AI or
17:25lose my job. And even when I use AI, I'm
17:28still seeing people lose their jobs. I
17:31just hate it. You know, I I hate having
17:34to leverage this technology rather than
17:37do my own my own thing and maybe ignore
17:40the technology, which obviously none of
17:43us can ignore at this point. There's a
17:45lot of people clearly that are very
17:47unhappy, resentful, disoriented. I
17:50imagine that group can't imagine. A lot
17:53of people are feeling extremely
17:55energized and I think the vice versa. So
17:58I think this is uh just like uh really
18:01interesting to see that some many people
18:03are very different from you. Uh there
18:06are many people that are having a really
18:07good time, a lot of people that are
18:08having a really bad time and it's just a
18:10a good uh way to remember that and give
18:13and basically have empathy for the other
18:15half especially the folks not doing as
18:17well because not everyone is as
18:19energized and we'll talk about like what
18:21that energized groups look group looks
18:23like but I think that's just an
18:25important point like a lot of people are
18:26having a really hard time. I I really
18:28agree with you Lenny and I think that
18:30given the pace of change of the
18:32technology. I I really do feel that too
18:35much of our focus is being spent on
18:38mastering the technology and trying to
18:40understand what the heck we should be
18:42doing with these new releases, new
18:44models, new capabilities,
18:47uh loop engineering, whatever the next
18:49thing will be next week. And we're not
18:52spending enough time and focus on our
18:55relationships at work, on seeing our
18:58colleagues, on supporting our peers in
19:02also being able to, you know, work with
19:04this technology thrive with it, succeed
19:07with it, and and get along and
19:10collaborate well with people who don't
19:13belong in the same bucket as us when it
19:14comes to how the technology is being uh
19:17perceived. So I personally would love to
19:19see all of us, especially energized
19:22people, take some of that energy and
19:26spend it on the people around you rather
19:29than the allin focus on whatever the
19:33next thing is with with AI. Let's go to
19:36the next takeaway. Let's talk about
19:38burnout and optimism.
19:40>> Yeah. So last year we looked at burnout
19:43and we looked at optimism and burnout
19:46was at a worrying level enough that we
19:50published further research on it and we
19:53introduced the ARM framework around how
19:56to deal with burnout and fight burnout.
19:59And this time around I'm not sure what I
20:03was expecting necessarily. Um, I think I
20:06was hoping that optimism would hold
20:09steady and that burnout wouldn't go up
20:12because supposedly with uh new abilities
20:16and even better models, perhaps you
20:18might think that we wouldn't have to
20:20work as hard or or or put as much energy
20:24into everything in a way that gets us
20:26burnt out. And yet, that's not what
20:29we're seeing. When it comes to
20:31significant burnout, burnout that's
20:33higher than moderate, we've gone up from
20:3644.7%
20:37in 2025 to 54.7%
20:41in 2026. So burnout is surging. More
20:46than half of our sample are feeling
20:48significantly burnt out at this point.
20:51And at the same time, optimism around
20:55the future of our roles and our careers
20:59is falling from 54.8% in 2025 to 48.7%
21:05in 2026.
21:07Bad news for the tech community. I don't
21:11know what to say about this other than
21:12that sucks. I don't like this. Uh I get
21:15it fully. Like I can completely
21:17understand why people are feeling this.
21:19Uh pretty worrisome. the the rise to
21:22like a massive jump like over 10%
21:25increase in burnout in just in a year.
21:28>> Yeah. And and it's interesting, you've
21:31had several guests on in the past who
21:34who've talked about things related to
21:36this. I loved your conversation with uh
21:39Jeff uh from RAMP.
21:43Um, and he shared that his worst burnout
21:48happened when velocity was lowest. So
21:51when you put effort into things that
21:52don't actually move,
21:54it doesn't feel good and that leads uh
21:57to burnout. And I think that maps for me
22:01with certain theories around burnout
22:04where stagnation leads to burnout. I
22:07think what's going on now is the
22:09opposite. These data showcase to you
22:11that as we ship faster than ever, as we
22:15move from shipping, you know, a couple
22:17of PRs a day to 30 PRs uh a day or or
22:22just doing so much more.
22:25In some ways, that might be fun, but we
22:27are burning out more than than we did
22:31before.
22:32And and we're not working any less hard,
22:36clearly, right? We we we're just taking
22:39on more stuff, more stuff, more
22:43prototypes, more PRDs,
22:46more PRs, more campaigns, more agents,
22:49more ads,
22:51>> and it's leading to more burnouts.
22:54>> Yeah. And that's actually one of the
22:55takeaways which we'll get to. Uh I don't
22:57know if you have this in this deck, but
22:58the a glimmer of hope that you pointed
23:01out is that enjoyment of work is
23:02actually still very high that it's
23:05burnout, but also people still at the
23:07same levels last year actually really
23:08enjoying their work. I think what we
23:10heard about enjoyment more than anything
23:12else which does give me hope is it's a
23:16couple of things I think um first of all
23:20people are able to bring out certain
23:24aspects of their identity and of their
23:27passions that they weren't able to
23:31before. You know, if if you're currently
23:33a PM by role, but you have a a designer
23:38within you or you're a marketer, but
23:40really inside you have that engineer
23:42that's been really wanting to to come
23:45out. Now, you can, right? Um, various
23:48tools that are available to us these
23:50days enable you to bring out those sides
23:53of your personality and those
23:55professional aspirations. And that's
23:57something that people are really
23:59enjoying. We're sort of escaping our
24:02classic swim lanes and a certain
24:06percentage of our roles are transforming
24:09into something else. And then the second
24:11thing that people are saying that
24:13they're finding enjoyable
24:16is that um they are able to manifest
24:22things that they've wanted to build that
24:24seemed impossible
24:26not very long ago, you know, um and that
24:30and that's whether it's in their main
24:32job or things that they're doing on the
24:34side, etc. Uh, so it is still for many
24:39people in tech an exciting time and an
24:42enjoyable time because we can do things
24:45that we never thought we'd be able to do
24:47because they're outside of our role
24:49because they're outside of what felt
24:51technically feasible or for other
24:54reasons. And this is another data point
24:56in the at the end of this you have a
24:58bunch of recommendations of what to do
25:00with all this information and that
25:02connects to something that you recommend
25:03later. But before we get there, uh, one
25:07of the other really interesting
25:08takeaways was around layoffs and how
25:10layoffs impact fear of layoffs impact
25:12burnout. Talk about that. We did ask
25:15people how worried they are about being
25:16laid off. And as you can see, 72% of our
25:21sample are worried to a certain extent
25:23about being laid off and 41.2%
25:27are at least moderately worried about
25:30being laid off. So it is clear that
25:33people are feeling that they might be uh
25:37cutting the branch that they're sitting
25:39on uh when it comes to using AI. Um
25:44people are reporting that their
25:45companies going all in on agents all in
25:49on these technologies and it does feel
25:52like they may be pushed out at some
25:55point. So the the worry level is is
25:59high. And I think if you're watching
26:02this, you're starting to notice that
26:04there's a ton of ambivalence right now
26:06in the community and a ton of uh clashes
26:10between emotions,
26:13feeling energized and enjoying my job
26:15and at the same time not being very
26:18optimistic about it and feeling tired
26:21from it and feeling worried that all of
26:23that enjoyment is going to end one day
26:26when the company I work for decides that
26:28I'm no longer needed. I feel a lot of
26:30people listening are probably feeling
26:32feeling all this and uh feeling heard.
26:35Uh so what's interesting is so far the
26:38kind of two takeaways. One is this
26:40bifurcation that's happening and these
26:41kind of archetypes that break that group
26:43a little bit further. At the same time
26:46just broadly burnout is increasing
26:49significantly. Optimism is declining.
26:52Talk about just those two kind of how
26:53you think about those two side by side.
26:55this idea of many people are much
26:57happier and having the best time of
26:59their life, many people are having a
27:00really bad time, but then broadly this
27:02optimism and burnout shift.
27:03>> I think there's something addictive
27:06about this technology. And I I'll speak
27:09for myself for a second and say that,
27:10you know, I've also [snorts] never had
27:13as much fun in my career as when using
27:15these technologies. And I've been using
27:17it for too many hours a day playing
27:21about and building. I think many of us,
27:24even those who aren't as energized about
27:26the technology, we still feel like we're
27:28in this fascinating technological
27:31playground.
27:32And it's very very hard to disengage
27:35from that technology and touch grass,
27:38read a book, whatever you love to do in
27:40in your spare time as well, Lenny. And
27:43so, yes, we're enjoying using this
27:46technology because it's opened up
27:47avenues that we've never had. But on the
27:49other hand, it's tiring. It's tiring to
27:52be constantly learning this technology
27:55and constantly trying to apply it in
27:56different contexts and constantly try to
27:59take on more than what your traditional
28:02role dictated because again there aren't
28:05any traditional roles anymore. We're
28:07just all builders uh these days whether
28:10that's an official thing within a
28:11company or not. And so I think that
28:15naturally
28:17this leads to these these clashing
28:20emotions, these exciting feelings around
28:24what's possible and how fun it is to use
28:27AI. And on the other hand, the
28:29realization that the reason this
28:32technology is so exciting is because
28:34it's so powerful. And because it's so
28:37powerful, it might be coming for us and
28:41to take our jobs, right? Did you Did you
28:45grow up on the Terminator movies? I keep
28:48thinking about the Terminator movies.
28:49>> I don't know about growing up, but yes,
28:50I I watched them early in my
28:52>> Oh my gosh, I love those movies and I
28:54just can't help but wonder, you know,
28:57when the robots are coming for us. So,
29:00uh, incredible technology
29:03might be coming for us at some point.
29:04Skynet Skynet is being built right now.
29:07>> Skynet is coming.
29:08>> So, there's a there's going to be a
29:09really great takeaway around this.
29:10Before we get to that one, actually, uh
29:12it's like every single one is so
29:13interesting and just like, man, this
29:14one's interesting, but they're all
29:15interesting. Let's get to the next
29:17takeaway around people's sense of their
29:19own career and whether they recommend
29:21it. This is a wild one for me because at
29:25least within the research community, I
29:27am known as the net promoter score
29:30hater. I even built a website dedicated
29:34to fighting NPS. NPS is the worst.com.
29:41>> It's an actual website. Fun fact, by the
29:43way, I I dislike NPS so much that I even
29:47brought the domain NPSthebest.com
29:51and I forward people from that to
29:54NPSW.com. That's how committed I am to
29:57fighting NPS. But but you and I were
30:01both genuinely curious about whether
30:05people who are, you know, already a few
30:08years in and they're maybe senior or
30:11above senior in their roles, what would
30:14they say to the person entering the
30:17industry now? Or what would they say to
30:19a person considering entering the
30:21industry now? would they recommend their
30:23role and and the tech industry more
30:26broadly to a friend or a relative of
30:29theirs?
30:31And the answer was shocking. Absolutely
30:36shocking. So, let me show you the data
30:40around this. And let me pre preface this
30:42with a quick explanation on NPS because
30:45it's not so easy to understand. NPS is
30:48this question around whether you would
30:50recommend a product or in this case your
30:53role to a friend or family member,
30:56right? And the scale goes from 0 to 10.
31:0110 being you would absolutely recommend
31:04your role and zero being you would
31:06absolutely not recommend your bowl. But
31:08then NPS as a whole when it's calculated
31:12goes from minus 100 which is absolutely
31:15terrible to plus 100. And the zero point
31:21represents neutrality. It represents
31:24that the sample as a whole is neutral.
31:27It's not a promoter to use NPS uh speak
31:33meaning I would promote my role to other
31:35people. I would recommend it to other
31:37people. and they're not detractors
31:39either, which means I'm not negative
31:41about my role. So that's the zero point.
31:44What you're seeing here
31:47is that no one is a promoter of their
31:51role in tech right now. Not even
31:54founders who are by far the happiest
31:58happy go-lucky people in in tech right
32:01now. Founders would not recommend their
32:04roles. Neither would people in sales or
32:07go to market. PMs, operations,
32:09engineering, and the worst of all,
32:12designers and researchers, my community
32:15are the least likely to recommend their
32:18role to other people coming into tech.
32:21It's sort of like saying, you know, I'm
32:24I'm kind of swimming in this pool. The
32:27water's kind of okay, but you shouldn't
32:29come into these these tech waters. They
32:32ain't for you.
32:34My god, what an interesting takeaway and
32:37and part of this research. So, just
32:38looking at this chart just as people are
32:40seeing this uh just like nobody
32:43everyone's like, "Nope, don't do what I
32:45do." It's interesting. Like, you know,
32:47we're going to talk about how founders
32:48are the happiest people in tech, but
32:50still it's like a very stressful hard
32:52job. Uh I'm surprised it's like I'm
32:55surprised it's the least like you would
32:57think many founders would be like this
32:59the craziest, gnarliest job. You should
33:02not do this sucks. But it's interesting.
33:03They're the least unhappy. They're the
33:05least least likely to recommend.
33:08>> There's a strong narrative right now
33:09that there's never been a better time to
33:12be a founder. You're a lot less reliant
33:14on other people to build.
33:17>> There's a ton of conversations around
33:19the soloreneur
33:21or the duoropreneur, right? It says a
33:24ton of conversations around how this is
33:27the era to be a founder and yet people
33:30who are currently in that role
33:33aren't very excited about recommending
33:34it to other people. There's also another
33:37layer to this which is as you go down in
33:42seniority you're less likely to
33:44recommend your role. So execs, VPs,
33:49maybe even directors are sort of more
33:53likely to recommend their role than
33:55people who are ICes. That's an
33:58interesting one. And I I've seen all
34:00sorts of explanations for that. Um, some
34:04one explanation for example is that at
34:07the top of the pyramid VPs are
34:10benefiting more from AI because they're
34:13getting all of this stream of
34:15information and knowledge being
34:17processed by AI making their jobs a lot
34:20easier. Whereas IC's are each scrambling
34:25to build all of these micro SAS products
34:28within the companies they work for. No
34:31one knows what anyone else is doing.
34:34There's a lot of duplicative
34:36work. It's a lot easier to build
34:40products these days, but a lot harder to
34:42maintain them. And so I think a lot of
34:45IC's are feeling like um there's a
34:47saying um full gas on neutral, right?
34:51Like you're pushing the pedal to the
34:52metal in your car, but you're in neutral
34:54gear. You're not going
34:56>> anywhere. I think that that's the
34:57feeling that a lot of IC's are having
34:59these days with these technologies. like
35:00I keep building these things but I'm not
35:02having the impact I want. Everyone else
35:05is building to there's a lot of
35:08confusion and it's just not worth it
35:11anymore.
35:13There's like this interesting also trend
35:14this, you know, there's this whole meme
35:16of uh don't become part of the permanent
35:19underclass. And I feel like there's like
35:21a few of these threads throughout this
35:23and here it's like okay I'm like a PM
35:25but like it's too late. Don't do this.
35:28like the door is closing for this role.
35:30Things are changing and that's it's not
35:32just PM, it's every role. Everyone's
35:34just like maybe should do something
35:35else.
35:36>> Yeah, I would agree with that and I
35:38think it's uh it's unfortunate.
35:41I
35:42I feel like in the past uh people were
35:47pushed to CS to learn to code. I know
35:52that these days for myself as a parent,
35:55you know, I don't know what to tell my
35:58kids to do anymore. I'm not sure what
36:01roles will be there for them in just a
36:04few years. These are very confusing
36:06times and I think tech employees are
36:08being that as well. Um they don't
36:10understand what's in line for them in
36:13the future. We certainly don't know to
36:15recommend to others like yeah you know
36:17in 5 years
36:19>> it's going to be great
36:21>> plastics
36:22>> like it's like no one knows it's just
36:24like just like zoom out for a moment.
36:26It's just it's so insane how much change
36:29we're living through right now. We would
36:31not have like this is not a result we
36:33would have seen a year or two ago. This
36:35is like so un unstable and uncertain and
36:41this is very new for for people in tech
36:44I think. Yeah,
36:46I like to use uh certain uh
36:51metaphors or or or sort of uh examples
36:54from you know other other uh areas in my
36:58life to explain these things. So, I
37:01believe you spoke in the past to uh
37:04Scott Woo, the co-founder and CEO of
37:06Cognition.
37:07>> Mhm. Devon
37:08>> and uh and yeah, the makers of Devon.
37:12And I recall that when he described the
37:15product, he talked about this this
37:18ladder that the product progressed on.
37:21It went from a high school CS student to
37:24a college intern to a junior engineer
37:27and at this point probably a senior or
37:29or a staff level engineer.
37:32I would say that we are all watching
37:36this technology
37:38climb the rungs of this ladder and
37:42advance and we feel like the technology
37:45is pulling those rungs from under us.
37:49And so the further we've gone up the
37:52ladder, the better we feel about our
37:55careers and the more stability we have.
37:57But the lower we are on that ladder,
38:00those rungs getting pulled under our
38:02feet. And so it makes lots of sense that
38:05we wouldn't recommend, you know, being
38:06on that ladder to anyone else because
38:08the rungs are disappearing.
38:11Man, there's so much to talk about with
38:13Elise, but let me point out a couple
38:15things as I look at this and then we can
38:17move on to the next takeaway. One is
38:18just like you know at the it's
38:20interesting that sales and PM are the
38:22least uh bad at don't get into this role
38:26like that's interesting I did not I
38:28wouldn't have expected that sales people
38:31not as scar scared about the future of
38:33the role PM interesting that's a big
38:34audience here not as scared uh as you
38:37said research design the most don't do
38:40this as a researcher you touched on this
38:42a little bit but I'm just curious how
38:44does it feel seeing this seeing this
38:46thought
38:48>> it's it's [snorts] tough. I think the
38:50research community has lived with uh a
38:54lot of concerns and trauma and anxiety
38:59over the years around having a seat at
39:01the table
39:03around finding meaning in our work
39:07around topics like the democratization
39:09of research and you know other people
39:12taking on our roles and eventually
39:16replacing us. It it definitely hurts a
39:18lot to see this. I think there's a very
39:21important role that research
39:24needs to play in the future of building
39:28products. I think that as this
39:31technology progressive, we should become
39:33even more thoughtful about what we build
39:37uh rather than than less thoughtful.
39:40I've heard the narratives around, you
39:42know, it's so cheap to build just build
39:44a thousand prototypes and see what
39:46works. I think you heard this in a
39:47recent conversation [snorts] uh as well.
39:51I would suggest that we're just going to
39:54get even more burnt out that way. I
39:56would not recommend that approach
39:58personally. I think research has a very
40:00and design has a very important part to
40:02play in raising the ceiling of what's
40:06possible. Um whereas AI has mostly
40:10lowered the bar, but it hasn't hasn't
40:13raised the ceiling. And I can also tell
40:14you that PMS being
40:18somewhat more likely to recommend their
40:20roles, even though they're they're not
40:22likely to recommend their roles, does
40:24actually make sense to me. It's more of
40:27a generalist role, I would claim, than
40:30some of these other roles. And this is
40:32the era of the generalist, if you would
40:34ask me, the the AI era. I'm the most
40:37surprised by sales and go to market. Uh
40:40because if we are indeed in some sort of
40:43SAS apocalypse, which I don't personally
40:45believe we are, but you know, it's been
40:46said, then I don't think uh sales and go
40:50to market people should be feeling too
40:51great about recommending their roles. Um
40:56but but maybe they believe that, you
40:58know, sales at this point cannot be done
41:02by AI.
41:03>> Yeah, there's just like so much I want
41:05to talk about here. Uh it's important to
41:06note this is a very important note that
41:08I'm going to say here. This doesn't mean
41:10design is going away or research is
41:12going away. This is just saying people
41:14in that role right now the narrative in
41:16their head is don't do this. Does not
41:20mean this is over because it could just
41:23be all this hype that we keep hearing
41:25like you know guilty of uh adding to it
41:28from various guests but you know this
41:31may this is just how people feel doesn't
41:32mean this what's happening. 100%. This
41:35entire unique project in tech is all
41:39about how people feel. It's not about
41:42the realities of working in tech. It's
41:44about people's emotions, about how
41:45people feel. Um, I really do think no
41:48one other than us has really captured
41:50these emotions in the way that we have.
41:52But yeah, that's a great reminder. This
41:54is not an image of reality. This is an
41:57image of how people are feeling in tech
41:59right now.
42:00>> Another quick data point that we want to
42:02make sure to communicate here. touched
42:03on this, but I don't know if you
42:04actually made this as as uh clear to
42:06people as it should be is uh so this is
42:09looking at would you get would you
42:11recommend people get into my role you we
42:13also asked just how are you feeling
42:14about your role like are you happy being
42:17in this role and the result there was
42:19I'm actually doing okay I'm happy I'm a
42:21PM I'm happy a designer researcher so as
42:23you said it's this kind of like the
42:24water's fine but don't get in talk a
42:26little bit about that just to make sure
42:27that point is clear
42:28>> when we asked people things like uh are
42:32they enjoying their their job as we
42:36talked about earlier. Uh most people
42:38said that overall yes, you know, they're
42:42they're having fun uh in in their role.
42:45I think what's going on here is that
42:48again people are lacking optimism
42:51compared to what they were, right?
42:53Optimism has gone down. So there's a big
42:56difference between how people are
42:58feeling now about their role given what
43:02they've achieved in their careers and
43:03where they are on the ladder compared to
43:08what they're extrapolating might happen
43:11in a year in 3 years in 5 years and what
43:15people are extrapolating might happen is
43:17clearly more negative than their current
43:20experiences. So yeah, I mean all in all
43:22we're doing okay as an industry. We're
43:25enjoying we're trying new things. We're
43:28experiencing uh a a a complete paradigm
43:31shift in how we work
43:35and that's fun for a while but it it
43:38seems like people feel like this is not
43:41going in the best of directions. And of
43:44course I hope we're all wrong. I hope
43:46we're all wrong.
43:48>> Me too. As a natural optimist, I'm like,
43:50it's gonna be all right. So, we'll see.
43:53I'll see.
43:53>> I'm a natural optimist, too. And I'm
43:55realizing like so far we haven't been in
43:57too much of optimism mode. So, I hope we
43:59manage to get to some of that as well.
44:02This episode is brought to you by
44:04Mercury, radically different banking,
44:06loved by over 300,000 entrepreneurs, and
44:09now with command. [music] I've been a
44:11customer of Mercury's for over 6 years.
44:13I have never once thought about leaving.
44:15Mercury is basically what happens when
44:17banking is built by product people, not
44:20by bankers. They make it so easy, dare I
44:23say fun, to send invoices, [music] move
44:25money around, set up virtual cards for
44:28folks on my team. Does your bank have an
44:30API, a terminal native CLI or an AI
44:33ready MCP server? I don't think so. And
44:36[music] just recently, they launched
44:37Command, a conversational interface
44:40built directly into Mercury, which acts
44:42[music] as your financial operator. I've
44:44been using Command to transfer money
44:46around to figure out what categories
44:47I've been spending the most money in,
44:49analyze my cash flows, and just today I
44:52used it to find out how much I've made
44:53from a specific sponsor over the past
44:55year. I just ask, "How much have I made
44:57from X [music] over the past year?" 10
44:59seconds later, I have an answer. It is
45:01so freaking cool. Visit mercury.com to
45:04learn more and apply online in minutes.
45:07Mercury is a fintech company, not an
45:08FDIC insured bank. banking services
45:10provided through Choice Financial Group
45:12and column NA members FDIC. Let's talk
45:15about takeaway four. Looked at
45:17productivity, looked at quality, look at
45:19just the kind of the upside of AI.
45:22This was a very interesting one to me.
45:25In last year's survey um that we did on
45:29the impact of AI specifically on
45:32productivity and quality, people were
45:34rather bullish about the technology and
45:37how much better it makes them not just
45:39more efficient but better. And this time
45:43around we saw some some interesting
45:47findings. Let me start with with this
45:50and this is sort of the initial view but
45:52it leads to something very different
45:54than I thought would happen. When we
45:55simply asked the question, how much
45:58better are you at your job thanks to AI?
46:02Unequivocally the answer was yes. You
46:04can see here that 97.2% 2% of people are
46:07saying that AI is making them better at
46:10their job and close to 50% of people are
46:13saying that AI is making them very much
46:16or extremely better at their job. So
46:20it's very clear that high level,
46:24however you interpret better, which
46:25we'll get to in a second, people feel
46:27like AI has had an impact. But when you
46:31dive deeper into what that means and you
46:35go below the surface level, as you can
46:37see here, we're doing better
46:42actually means something rather
46:44different than what you might think. So
46:46when I heard people say they're doing
46:48better, I thought they meant that
46:50literally. I thought the quality of
46:53their work was better, that AI was
46:55contributing to better quality of work.
46:57In fact, it's the exact opposite. What
47:01we heard from people were two concerning
47:03things. The first is that I can do more
47:07faster, but not better. So again, AI has
47:12sort of lowered the floor of what's
47:15possible and and enabled people to put
47:17out a lot more work than they ever have
47:20before, whatever that work is, whether
47:22it's uh PRs, PRDs, research projects,
47:26prototypes, but they're not the output
47:30isn't of higher quality.
47:34And perhaps the even more concerning
47:36effect which we heard about the the
47:38deeper cost of all of this is the cost
47:42of
47:44using AI on thinking and judgment.
47:48People essentially told us my brain is
47:50rotting my work feels worse. And this
47:55this is related to phenomenon I've heard
48:00about recently like um is it cognitive
48:02rot I think is what what I've heard.
48:07Basically this phenomenon where you sort
48:10of see the initial output of an AI model
48:16and you sort of just accept it. You
48:18don't apply your judgment to it. You
48:20don't apply thinking to it and you
48:23slowly let your your mind, your
48:27involvement in your work, your agency
48:31sort of collapse into this rotting
48:34state, so to speak.
48:36>> This Yeah, this is such an interesting
48:38takeaway just uh like this is a positive
48:40there's a positive note to this. As you
48:42said, I'm looking at the numbers here.
48:4482% of people say AI is making them at
48:46least moderately better at their job and
48:48nearly 50% say very much or extremely
48:51better which may correlate with that
48:52original just this bifurcation like the
48:54people that feel like AI is making them
48:56better at their job probably the people
48:57feel like most energized
48:59>> so whether it's true or not so I think
49:01there's a couple happy points that's one
49:03people feel AI is making them better
49:04they like they feel that so that's you
49:06know that's real even if it may not be
49:09and at the same time I think they're
49:10self-aware the quality is an issue And
49:14also I'm just maybe my I'm atrophying in
49:17my ability to say code or write strategy
49:21docs. So at least people are self-aware
49:23that this is happening. Not just I'm
49:24killing it. All my stuff is higher
49:26quality and it's much better than ever.
49:28>> They are self-aware but you know among
49:30other things I'm a certified coach and I
49:33would suggest that beyond being
49:34self-aware we need to act
49:36>> on this. There are several explanations
49:38I could give to this phenomenon. One of
49:40them might be that the honeymoon period
49:43with AI is over. You know, we thought it
49:46was perfect and, you know, the the best
49:50thing that ever happened to us since
49:52sliced bread or whatever typical
49:54Americans say. Um, but we've realized
49:58it's not quite there yet. And yes, the
50:00models are the worst they're ever going
50:01to be. They're only getting better, but
50:03there's still so much they need to get
50:06better before we can really say, "Yeah,
50:07like my work is just so much better."
50:09and I don't need to be as involved with
50:12it. Uh you had Simon Willis on on this
50:16show talking about how people are
50:19worried about skill atrophy. Um that
50:22they're not learning enough and that we
50:25need to be consciously resistant to this
50:29and not let ourselves cognitively rot.
50:33And I think this data suggests that
50:35we're losing that fight a little bit. We
50:37need to put to put a little bit more
50:39energy into it into fighting this
50:41cognitive rot that that we're feeling
50:45and to make sure that as the bar
50:49continues to rise and this technology
50:51continues to evolve that we're meeting
50:53that with deliberate practice,
50:56deliberate thinking, deliberate
50:58judgment. People are just so bad at
51:00that. Just like if I have an easy
51:01button, I will press it. I no I
51:03shouldn't, but I'm just going to going
51:05to eat those Doritos. So that's the
51:06challenge I think for people is just
51:08like because you could do it the easy
51:09way. It's so hard to resist. And that
51:11connects again to this ladder that's
51:13being raised that people are just
51:15getting into this stuff. It's just they
51:17have to make themselves do it the hard
51:18way to learn how to do it for real. And
51:20I'm really curious how that plays out
51:22with junior folks in the market learning
51:25how to actually code, how to do writing,
51:27how to write a PRD from scratch with
51:28your fingers versus talking to Clott
51:30about it.
51:31>> Yeah, there's so much to say here. I
51:33think um the the more I think people
51:35miss feeling smart sometimes, you know.
51:39I think I I thought
51:41>> once upon a time about myself that I'm
51:42I'm smart and then you encounter these
51:45these models, most recently Fable Five
51:48that just came out for the second time
51:50and you don't feel that smart anymore
51:52and you sort of lose this sense of
51:54self-efficacy that that you used to
51:57have. And and the second thing to say
52:00about this is that every time you you
52:03not the AI solve a problem and get over
52:06a barrier, it increases it raises your
52:10baseline of self-efficacy, of
52:12self-confidence, of self-belief. And
52:15every time you offload that to your
52:18favorite AI model, you're lowering that
52:21baseline and your thinking and your
52:23judgment is watching. And that's a
52:24serious problem. So maybe to just wrap
52:27up this takeaway, the sentence used in
52:29this report I love. I'll just read the
52:31productivity gains are real, but the
52:32quality of the work and the sharpness of
52:34the person producing it are taking a
52:36hit.
52:37>> Absolutely.
52:39And well, maybe we'll get to this in the
52:40end, but we should keep a close eye on
52:43this and really try to work on our own
52:46self-efficacy.
52:47>> We need like evals for people like are
52:48humans becoming
52:49>> eval for people. Let's let's do that.
52:52>> Your smartness. Okay, let's talk about
52:53takeaway number five, which I also
52:56thought was extremely interesting. My
52:57god, so much so much interesting in
53:00this. Let's talk about the fifth
53:02takeaway. Yeah. So, I think the dominant
53:06narrative right now in in our community
53:08is is the fear of being replaced. And as
53:10we talked about, a lot of people in tech
53:14do fear and are worried that they're
53:16going to be replaced, that they're going
53:17to be laid off. And so it would make
53:21sense that if we had a question and one
53:25of the options to respond to that
53:28question were, I'm worried about losing
53:30my job to AI, that it would rank fairly
53:33high. But when we look at this slide and
53:36when we look at when people are asked,
53:38you know, what are your concerns and
53:41what's bothering you right now? What
53:43what are you afraid of in tech? Losing
53:46my job to AI is actually second to last
53:50on that list. And instead, what we saw
53:53rise up to the top is the expectation to
53:56do more for the same pay. It's almost as
53:59if we we're all feeling like we're being
54:01squeezed
54:02and squeeze and squeeze some more to
54:06accomplish more things in less time for
54:09the same pay. And so that fear of being
54:13overworked is by far the the dominant
54:16fear alongside the second item in this
54:20list which is that the pace is becoming
54:23unsustainable.
54:24And by pace I mean both the pace of the
54:28work and the expectation of work
54:31velocity just how much you have to put
54:32out a day every single day and the pace
54:37of change of the technology which
54:40requires you to invest a significant
54:42amount of time in learning the new model
54:45the new framework the new way to use the
54:49technology which of course takes away
54:52from your core focused working time. So,
54:56it's sort of an evil downward spiral
55:00that really prevents you from uh feeling
55:03like you're able to accomplish anything
55:05of of meaning. You just feel very
55:08overworked, very tired.
55:10I'm confident everybody listening uh
55:12understands exactly what you're
55:14describing here. Uh obviously connects
55:17to that burnout point we made earlier.
55:19Uh I'll close this takeaway with uh
55:21another quick read from your post.
55:22You're such an amazing writer, Nom.
55:25>> I love read I love reading your uh your
55:27summaries of of what you learned. So,
55:29I'll just read this. The speed AI
55:31unlocked got plowed straight back into
55:33expectations. Every game becomes the new
55:35baseline and the people expected to hit
55:37it are running out of room to breathe.
55:40>> When are we going to get to the
55:41optimistic part of this uh of this
55:43episode?
55:46>> Maybe we'll find something. Okay, let's
55:48talk about 6, which has a little bit of
55:50optimism.
55:51>> Oh, yeah. Okay. a lot of ambival
55:53ambivalence is another way you put it.
55:56>> Yeah, I mentioned this earlier. The
55:58industry is definitely in a state of
56:01ambivalence. We we try to capture all of
56:04the emotions that people are feeling at
56:07work right now. And I am happy to report
56:11as you can see here that the top two
56:13emotions are positive emotions,
56:17curiosity and excitement. And that's
56:20wonderful. that is optimistic. I'm glad
56:23that's the case. People are sharing that
56:27they're excited about the technology,
56:30about the the capabilities, about what
56:32they're able to achieve. They're feeling
56:35curious about what the future might
56:38bring and what else they could do as
56:40these models and systems continue to
56:43evolve. So, that's wonderful.
56:46But as you go down this list, you can
56:48see that there's this very interesting
56:50mix of positive, negative, and neutral
56:53emotions. People are also feeling
56:55overwhelmed. People are feeling
56:57conflicted. People are feeling relieved
57:01in certain ways um about, you know, what
57:05AI has taken from them in a in a good
57:07sense, you know, what they've been able
57:09to offload to AI. And then people are
57:12feeling tired,
57:15burnt out, just working hard and hard as
57:18we said, uneasy, anxious, and then
57:21hopeful. So, as you can see, there is a
57:25wide variety of of emotions here. And
57:29it's uh it's really tough to
57:33to figure out what exactly is going on.
57:36But one thing I I definitely wanted to
57:38mention which uh Nikil Singal mentioned
57:41in an episode with you, I know he's a
57:43friend of the pod and such an incredible
57:46leader in in product and in general in
57:49the tech community. He talked about
57:52smiling exhaustion.
57:55And I just love that term because I
57:58think it captures this ambivalence. the
58:00the old type of emotion, the old type of
58:04burnout was just completely grim, you
58:08know, it it's it it was all about these
58:11negative emotions that come from
58:13disengagement and exhaustion. Whereas
58:16now, when Kill talks about signing
58:19exhaustion, it's about people almost
58:22feeling reborn on the one hand. You
58:26know, I'm shipping again. I'm building.
58:29I'm creating these incredible things
58:31with AI. There's never been a more
58:33exciting time, but there's no off
58:37switch. There's no off switch. And so,
58:40the tempo is absolutely brutal. The
58:44rules keep rewriting themselves every
58:47single day. It's a very relentless pace.
58:52And so, we're smiling through that
58:55exhaustion that that we're feeling. And
58:57I think that captures the ambivalence
58:58more than than anything really.
59:00>> I think it's also good to kind of come
59:02back to one of the core points you made
59:04earlier that half of people are actually
59:07feeling very like they are working very
59:09hard probably exhausted at times but
59:12they're also just like really into
59:14they're going all in uh feeling good
59:17energized about where things are going.
59:18So I think it's important to always
59:19remember like kind of come back to that
59:21as we share all these negative elements.
59:24There's actually a large group of people
59:26that are actually loving this time and
59:28doing super well. There are I think more
59:32than anything what I would like to
59:33highlight here is that you know we tend
59:36to think of things way too often in
59:39black and white
59:40>> Yeah.
59:41>> terms. We see people as either people
59:44who are just full of hype or people who
59:47are is it doomers would that be the
59:49right
59:50>> you know or haters or whatever you want
59:53to call it. What we're seeing here is
59:56that as always in life, if you ask me,
59:59things aren't as simple. Um, behind that
1:00:02facade, behind the theatrics that Elena
1:00:05Bernard discussed,
1:00:07>> we all have both the height person
1:00:10within us and the doomer to a certain
1:00:13extent. And the only difference between
1:00:15us is that we have different amounts of
1:00:18each. But we're not just one or the
1:00:22other. And there really is a ton of
1:00:23ambivalence and and I just want people
1:00:26to know that it's okay, you know, it's
1:00:28okay to be excited and curious, but also
1:00:30feel overwhelmed. It's okay to feel
1:00:34relieved and tired. It's okay to feel
1:00:36resentful. Even if all in all you feel
1:00:39positive about the technology, those
1:00:42emotions can coexist within us. It's
1:00:44very natural. Don't let this narrative
1:00:47that you're either or um to to dominate
1:00:50and to convince you that anything's
1:00:53wrong with how you feel because really
1:00:54all of these emotions are very
1:00:56understandable given the times we're in.
1:00:59Such an important point. I love that you
1:01:01said that.
1:01:02Okay. So the next takeaway what we did
1:01:05here is when we were looking at this
1:01:06bifurcation that you found, we looked at
1:01:09just which functions are having the best
1:01:11time and worst time. And uh we had a
1:01:14hint at this earlier uh and it turns out
1:01:16there's a couple that are just overall
1:01:18having the worst time right now and
1:01:20feeling the least happy. So let's get
1:01:22into that.
1:01:24Yeah. So definitely definitely and we
1:01:26saw this last year by the way as well.
1:01:29We are seeing that designers and
1:01:31researchers are the most negative. And
1:01:34just to clarify what I mean by that,
1:01:35what you can see in this slide is that
1:01:38when it comes to AI identity shift,
1:01:41designers and researchers are the
1:01:43highest in feeling either destabilized
1:01:45or diminished. When it comes to emotions
1:01:48like feeling tired, overwhelmed, or
1:01:51anxious, researchers and designers uh
1:01:54lead the pack in in a negative way. Of
1:01:57course, when it comes to worries about
1:02:00losing Guan's job to AI, again,
1:02:02researchers and designers are pretty
1:02:04much as high as one can be. And then
1:02:08finally, when it comes to these NPS-
1:02:11like scores of whether people would
1:02:13recommend their role to a friend or to a
1:02:15colleague, researchers and designers are
1:02:17the least likely to recommend entering
1:02:20their role at this point.
1:02:21>> We talked about this earlier. I don't
1:02:22know how much more there's to add here.
1:02:24Basically, it's again, it's not that
1:02:27this function is in big trouble
1:02:29necessarily. It's this is how people in
1:02:30that function are feeling right now. You
1:02:33know, I'm not objective because I'm part
1:02:35of the research community and the design
1:02:37community. Um, if I may, I just want to
1:02:41tell people that in this community that
1:02:45um, I truly and honestly believe that
1:02:48there's never been a more important time
1:02:52for our roles to manifest in the
1:02:54products that we build. Um, you've
1:02:56talked about this with many people,
1:02:58Lenny, in in previous interviews about
1:03:00concept like taste, about concepts like
1:03:04craft,
1:03:05about concepts like quality. I'm
1:03:08thinking of people like Kari Line's CEO
1:03:12who talked so beautifully about the
1:03:14importance of quality and how you build
1:03:17products. I'm thinking about people like
1:03:20uh uh like um
1:03:23>> Katie Dill comes to mind. She had a we
1:03:25had a whole conversation around taste
1:03:26and beauty and value. Katy D uh recently
1:03:31uh Jenny Wen who leads design uh for uh
1:03:34Claude at Anthropic and so many other
1:03:37people who talk about the importance of
1:03:40taste of judgment of design. So I I
1:03:45honestly hope that designers and
1:03:47researchers make a comeback when it
1:03:49comes to how they feel about this
1:03:50industry because the industry needs
1:03:53them. The industry needs us. And so I
1:03:56hope we see a change next year and this
1:04:00is the call to all designers and
1:04:01researchers to
1:04:04get in there and do our thing because
1:04:06it's incredibly valuable at this era.
1:04:08>> Yeah, I've been I've had a few
1:04:09conversations recently about just design
1:04:11and in particular just why why is design
1:04:14not a huge value add and differentiator
1:04:17today knowing AI is just so bad at it.
1:04:19We just had the head of the Codex app on
1:04:21talking about why AI like AI is not good
1:04:24at design and you could clearly tell AI
1:04:26designed this thing and you would think
1:04:28this was the moment for design to become
1:04:30such a value add and differentiator for
1:04:33companies seeing all this AI slop that's
1:04:35being produced. So I'm optimistic
1:04:37that'll emerge as people are like I'm so
1:04:39sick and tired of all these generic
1:04:40things. I I want to actually build
1:04:42something great. Yeah. And moreover, if
1:04:44I recall correctly, you you in that
1:04:46conversation talked about how some of
1:04:48the issues around AI doing design are
1:04:51perhaps tractable problems, but some of
1:04:53them are even intractable I I would say
1:04:57or definitely much much harder to solve
1:05:00for. It's still the case that creating
1:05:03incredible
1:05:04novel creative experiences
1:05:08not something that this technology is
1:05:10quite capable of. So, we'll see what
1:05:11happens, but I I really do hope that the
1:05:14designer research community turn around
1:05:16and change their minds about the
1:05:18industry next year.
1:05:20>> Yeah. Man, I can't wait to do next year
1:05:22already. What's it going to be? But if
1:05:25you actually look at the losing my job
1:05:27to AI bucket, there's actually a role
1:05:29that is even more worried about losing
1:05:30their job. And I believe that is data
1:05:32analytics. Our shades are kind of uh
1:05:35very similar to keep it to keep the
1:05:37message clear, but that's really
1:05:38interesting. the data anal data analysts
1:05:40basically are the most worried about
1:05:41losing their job.
1:05:43>> I think we're we're seeing uh incredible
1:05:47capabilities coming out.
1:05:49>> Makes sense.
1:05:50>> Yeah. With the different the different
1:05:51models out there to to run data
1:05:54analysis. I'm seeing this in the
1:05:56research realm as well. I would suggest
1:06:00that we're also seeing incredible
1:06:01advancements in coding. So I was
1:06:03expecting engineering to be higher in in
1:06:06certain places here and and it's not. Um
1:06:10but again this isn't about the objective
1:06:13reality of how far AI has come in terms
1:06:17of its capabilities doing each of these
1:06:19roles. This is about how people are
1:06:21feeling in these roles and there isn't a
1:06:24onetoone correlation uh between between
1:06:26those things.
1:06:28>> Speaking of that, let's talk about who
1:06:30the happiest people are. Who's doing
1:06:32great? Who's doing best?
1:06:34>> So, the happiest people in tech. Um,
1:06:37this is something that did not change
1:06:40between last year's survey and this
1:06:42year's, it's still founders and people
1:06:46who are working in smaller companies.
1:06:50And I do feel the need to say that when
1:06:52it comes to founders to your point
1:06:54earlier being not that I know but being
1:06:56a founder is an incredibly challenging
1:07:00all-encompassing
1:07:01job and there is a certain selection
1:07:05bias here in the sense that these are
1:07:07people who are currently in the active
1:07:11role of a founder in a startup that is
1:07:14is still running rather than founders
1:07:18who had to shut down their startup.
1:07:20purple, whatever it might be. So, I just
1:07:22want to make that clear around who we're
1:07:24looking at. But but yes, when we look at
1:07:27founders and when we look at in this
1:07:29case, you're looking at um company size
1:07:32as well, you can see that whether it's
1:07:36optimism, whether it's uh uh burnout,
1:07:40layoff worry, or whether people would
1:07:42recommend their job, uh you're just
1:07:45seeing these these uh increases or
1:07:48decreases depending on on the question.
1:07:50It's very clear that founders are top of
1:07:55nearly every measure. They're 71%
1:07:58optimistic. Uh they're enjoying their
1:08:01jobs the most. They have the lowest
1:08:04burnout, the lowest layoff worry, the
1:08:06most AI excitement. It's a it's a
1:08:09medium-sized effect compared to other
1:08:12effects we saw. Um but very significant.
1:08:16And and that said, you know, if we talk
1:08:18about founders, even having that
1:08:21ownership and even having that agency
1:08:24has its limits, um founders are still uh
1:08:28you know, many founders are still
1:08:30moderately or at least moderately burnt
1:08:32out. 47% of founders that would be um
1:08:36and as we said earlier, even founders
1:08:39aren't very likely to recommend being a
1:08:42founder to other people right now, which
1:08:44is very interesting. I think it's also
1:08:46important to note uh there's definitely
1:08:48this as you said this kind of meme of
1:08:50there's never been a better time to
1:08:52build. People may see this and be like I
1:08:55got to start a company. It's the best
1:08:56place to be. It may be this moment
1:08:59currently where people feeling most
1:09:00excited about it and then as the bubble
1:09:03you know I don't know if it's a bubble
1:09:04but as things start to get harder there
1:09:06may be a shift away from that. So this
1:09:09so far interesting that we've seen two
1:09:11years in a row founders are the happiest
1:09:13but I wonder if that changes if the
1:09:14market changes if some you know that
1:09:16kind of thing. So this isn't saying
1:09:18you're going to be happy if you become a
1:09:20founder but it is really interesting
1:09:22that two years in a row of all of the
1:09:24like the most uh clear one of the
1:09:27clearest takeaways is just founders seem
1:09:29to be happiest across every dimension.
1:09:31>> Yeah. And then company size uh as well.
1:09:35Again, uh optimism goes down if you're
1:09:39working for a larger company. Um you get
1:09:41more burnt out if you're working for a
1:09:43larger company. You're more worried
1:09:45about layoffs if you're working for a
1:09:47larger company. And you're less likely
1:09:48to recommend your role if you're working
1:09:50for a larger company. So that also
1:09:54remained consistent. And if we look just
1:09:59at burnout specifically, then you can
1:10:02see very clearly here that burnout
1:10:04climbs in very significant ways
1:10:07in between working for a small 1 to 10
1:10:10person startup and working for a 5,000
1:10:13or 10,000 person enterprise company.
1:10:18>> It's so wild. It's wild how linear this
1:10:20is. like there's no bump and then those
1:10:23two at the end there as you point out
1:10:24here in the chart they're statistic
1:10:26>> it's within margin of error yeah
1:10:28>> it's like wild here just like only goes
1:10:30up there's not like a sweet spot and
1:10:32then it comes down again and then if you
1:10:33go back to the previous chart exactly
1:10:35the same thing it's just like very
1:10:37linear everything goes up uh and gets
1:10:40worse as your company size grows that's
1:10:43wild
1:10:44>> absolutely I would say one thing to
1:10:46highlight here like if you look at
1:10:47burnout for example is that even the the
1:10:51the lowest is still fairly high, right?
1:10:55So, I would say people are feeling
1:10:57somewhat less burnt out, somewhat
1:10:59healthier working for a startup. But I I
1:11:01would view these levels of burnout, of
1:11:05worry as uh as worrisome. I'm worried
1:11:10about the level of worry right now in in
1:11:13tech. It is not a good vibe. You know,
1:11:17you mentioned how we met a decade ago at
1:11:19Airbnb. I do feel like the vibe was very
1:11:22different then in in tech. So, I am a
1:11:25bit concerned about the numbers that
1:11:26we're seeing here. But yes, from a
1:11:28relative sense, a relative point of
1:11:31view, um perhaps one should consider
1:11:36doing a shift and starting your own
1:11:38thing or working for a smaller company.
1:11:40>> Yeah, exactly. That'll be one of the
1:11:41takeaways. You know what would be
1:11:42awesome is if we do one of these for non
1:11:44tech. I'm really curious just like you
1:11:46think people are scared and worried in
1:11:49tech. I'm so curious how that would
1:11:51compare to you know like say the blue
1:11:52collar folks or other functions like I
1:11:56don't know if we have the audience to do
1:11:57that but that'd be really interesting. I
1:11:59mean I think my uh HVAC technician is
1:12:02feeling pretty great about his career
1:12:04right now burning hot in Florida.
1:12:07Everyone needs air conditioning. That's
1:12:09not going to change anytime soon. U take
1:12:13a hack. Yeah. To start a company,
1:12:15>> we should do that.
1:12:16>> Start a filming business. Oh, man.
1:12:19>> So, what I love about this doing the
1:12:21survey again is we're starting to find
1:12:23things that are consistently true. And
1:12:25the next takeaway is another really good
1:12:27example that we saw last year, too, uh,
1:12:29around managers. Let's get into that.
1:12:33Yeah. So, I mean, perhaps people are
1:12:38aware of this. Managers matter a ton.
1:12:42who your manager is probably matters
1:12:44more than most other characteristics of
1:12:47your role. And we found this very
1:12:49explicitly last year and this year again
1:12:54um manager effectiveness. So we
1:12:56basically asked people how effective is
1:12:59your manager? How would you rate your
1:13:01manager? And what you can see here is
1:13:04that the more effective your manager is,
1:13:07the less burnt out you feel and the more
1:13:11you enjoy your job. And this is a
1:13:14massive massive effect that we're
1:13:18seeing. So if you have an extremely
1:13:22effective manager, for example, you're
1:13:24reporting around 65% higher job
1:13:28enjoyment and dramatically lower
1:13:31burnout. The problem is
1:13:35is the the supply. um only about 25% of
1:13:40the sample rate their manager as highly
1:13:44effective and 36% of the sample rated
1:13:47their managers as ineffective which by
1:13:50the way is a very small change from from
1:13:54last year. So we're seeing managers be
1:13:58rated not so well and at the same time
1:14:02we're seeing that when managers are
1:14:04effective which is clearly uh not often
1:14:07enough the impact that has on people's
1:14:11enjoyment burnout and other aspects as
1:14:14well is incredible.
1:14:16That's wild. like like I think we're
1:14:19adding new uh insight into the world
1:14:22here just finding this I like I've never
1:14:24never seen this before just how
1:14:26impactful a manager is that we always
1:14:28hear anecdotes we always hear like you
1:14:30know people only leave they don't leave
1:14:32jobs they leave their manager but this
1:14:34is wild just how much that impacts you
1:14:36and clearly one of the takeaways might
1:14:38be go if you're having a bad time find a
1:14:40better manager as you said easier said
1:14:42than done
1:14:43>> yeah but there's a couple of things here
1:14:45I want to I want to also note before we
1:14:46move on. First of all, we are in the era
1:14:49of the great flattening. We're in the
1:14:52era of founder mode. People have more
1:14:55direct reports now than ever before. Um,
1:14:58we're trying to keep things as flat as
1:15:00possible and remove hierarchy. I'm
1:15:02concerned about this. I think, um,
1:15:04obviously there are good arguments for
1:15:07flatness and lack of hierarchy. Um but I
1:15:12am curious to see where this goes given
1:15:15the importance of who your manager is on
1:15:18your well-being. And then the second
1:15:20thing is we talked earlier about being
1:15:23overworked about this squeeze that AI
1:15:26has on us. Who's the person more than
1:15:29anyone else who kind of manages that
1:15:32squeeze? It's your manager, right? It's
1:15:34your manager who protects you. It's your
1:15:37manager who in many ways almost dictates
1:15:40uh how much you may be overworked or
1:15:43not. And so yeah, this this matters more
1:15:47than ever before. And I'm definitely
1:15:48concerned that some of the negative
1:15:50phenomenon we're seeing are a function
1:15:54of what we're doing to managers in tech
1:15:56and how we're building out our
1:15:58organizations. Such an interesting
1:16:00point. we whether we do it now or next
1:16:03time just looking at these metrics for
1:16:05managers like how much more burned out
1:16:06are they how much less happy and
1:16:08optimistic are they would be really
1:16:09interesting there's a couple notes here
1:16:11in in the actual post that I want to
1:16:13highlight that I thought are really
1:16:14interesting one is just core takeaway
1:16:16here is the biggest lever you have to
1:16:18increase retention in your company is
1:16:21improve your managers put people on with
1:16:24connect people with better managers
1:16:26basically so interesting that's like a
1:16:27thing you can do obviously easier said
1:16:29than done but that's the biggest lever
1:16:30of things you can change. And then the
1:16:32other thing you noted is the worst rated
1:16:34managers cluster in data analytics and
1:16:39design poor design. Yes. So, and I'm not
1:16:43sure there's a clear explanation uh for
1:16:47that. Um I think one possible
1:16:51explanation is that again you know
1:16:54managers are just people too and so
1:16:58design managers and data and analytic
1:17:00managers are are
1:17:03in design and in analytics they are
1:17:06suffering clearly based on what we saw
1:17:10and I think it's hard to separate how
1:17:13you feel in tech from your job as a
1:17:18manager, it's hard not to put some of
1:17:21your negative energy onto your team and
1:17:24the people you work with. And and so I I
1:17:27have a lot I have a lot of empathy for
1:17:30design managers, research managers, data
1:17:32and analytics managers, and other
1:17:34managers in tech who are having a hard
1:17:37time right now and they're not feeling
1:17:38energized. And unfortunately, some of
1:17:41that is being pushed on to their direct
1:17:45reports.
1:17:46And then [clears throat]
1:17:47>> and then I very much appreciate in some
1:17:50of the companies I've worked for having
1:17:52the opportunity to go through manager
1:17:55training which is something that's still
1:17:57all too rare I feel within
1:18:00organizations. If you're listening to
1:18:01this and you're a leader in an
1:18:04organization, if you're seuite at a
1:18:06company, please invest more in your
1:18:09managers. And if you're looking for a
1:18:11job right now, consider carefully who
1:18:14your manager might be because probably
1:18:18nothing will have more impact on your
1:18:20well-being than who that person ends up
1:18:22being.
1:18:23And to build on how uh difficult things
1:18:26are for companies right now, the level
1:18:28of poaching and offers throwing out
1:18:31thrown out at the best people, just
1:18:33imagine how sucky it is to find a great
1:18:37manager, train people to become awesome
1:18:39managers, and then they get, you know,
1:18:40>> only to lose them
1:18:41>> only to lose them to, you know, the
1:18:43labs, people with infinite money. So,
1:18:45it's just extra hard. which is a good
1:18:47segue to our final takeaway, which is a
1:18:49really good summary of just what is what
1:18:51people are feeling right now.
1:18:53>> Yeah, it is the the wildest it's ever
1:18:56been. We asked people to describe the
1:19:00state of the tech industry right now. We
1:19:03got thousands of responses. These were
1:19:07essentially open-ended responses, and we
1:19:09just sort of clustered those
1:19:11descriptions into this uh this word
1:19:14cloud. I still love wordcloud. Some
1:19:17researchers are sort of anti-wordcloud.
1:19:19I love I'm not gonna apologize for the
1:19:22word cloud.
1:19:23>> No, this is amazing.
1:19:24>> It's like such a clear. Look at this.
1:19:25Just look at this. This is exactly what
1:19:27it feels like.
1:19:28>> Yeah. We're seeing change. We're seeing
1:19:30chaos. We're seeing speed. We're seeing
1:19:33excitement, flux, hype, a lack of
1:19:37stability, a potential bubble, but also
1:19:40crazy opportunities. Everything's
1:19:43evolving. Everything's unstable. It
1:19:45comes with costs. It's better. It's
1:19:48worse. It's huge. It's confusing.
1:19:52All smooshed together into a more
1:19:56chaotic industry than it's ever been.
1:19:59Um, I don't understand anything about
1:20:02baseball or maybe it's cricket, but we
1:20:04got this quote. Do you understand
1:20:06baseball? Does
1:20:08>> hopefully if you're watching you
1:20:09understand these things, but we got this
1:20:11quote. We're in the second inning of a
1:20:14massive shift. No one knows how it will
1:20:17end, but all you can do is keep taking
1:20:20at bats. Is that a cricket thing? A
1:20:22baseball.
1:20:23>> That's baseball. I think it's I don't
1:20:24know how many innings. I don't know if
1:20:25there's innings in cricket, but it's
1:20:27definitely baseball. Uh it works for
1:20:29baseball at least.
1:20:30>> Yeah. Basically, it's just saying keep
1:20:31trying. Just keep taking shots. Shots on
1:20:34goal. A lot of metaphors.
1:20:36>> Yeah. So many good quotes. Another one I
1:20:38I see here which was which was good one.
1:20:41Tech is manic. Half out of touch,
1:20:45clinging to the bandwagon, porbing into
1:20:47the overhype. The other half are
1:20:49exhausted by the first half that came
1:20:52from a senior PM. Um yeah, fascinating
1:20:56stuff. When we look at this overall,
1:20:59it's crazy how we see again this
1:21:02bifocation that 37%
1:21:05of the words used when we ran sentiment
1:21:09analysis were positive, 37%
1:21:13were negative, and the the remaining 26%
1:21:17or so were neutral. Basically, an even
1:21:21split.
1:21:23We're all seeing the same thing, but
1:21:25half of us find it thrilling, the other
1:21:28half find it terrifying. Um, I was I
1:21:32was, you know, rather taken aback when I
1:21:35when I saw these data. It's pretty wild
1:21:37how equally uh uh divided we are.
1:21:42>> Yeah, man. And uh I think you used this
1:21:46quote earlier that this is the most
1:21:47normal it's ever going to be.
1:21:49>> I mean, I I don't know. Um, the prophecy
1:21:51was given to the fool, so I'm not going
1:21:53to try and be a be a prophet. But, um,
1:21:56yes, I I doubt it's going to get, you
1:21:58know, more normal and normalized
1:22:00statistically than than this. So,
1:22:04we'll we'll wait and see, I guess.
1:22:07>> Yeah. Like, if we zoom out, we're living
1:22:10through history right now. The amount of
1:22:13change and like this is unprecedented.
1:22:15You know, the industrial revolution is
1:22:16always the metaphor people use, which is
1:22:18accurate. uh maybe crazy or maybe not,
1:22:21but it's that's the interesting part.
1:22:22Like what an interesting time to be
1:22:24living through. Elon had this really
1:22:26interesting way of describing it. Like
1:22:28we're definitely living in a simulation
1:22:29if we're if we're alive right now when
1:22:32we're about to start building data
1:22:33centers in space and AI is going to be
1:22:36as smart as human brains. Like what are
1:22:37the chances you are alive at that
1:22:40moment? Uh which is probably okay. Maybe
1:22:42some higher intelligence is simulating
1:22:44the most interesting time in history.
1:22:47Uh, I don't know if that's a good rap,
1:22:49but
1:22:50>> listen listen, I have a lot of respect
1:22:54for for uh Elon. Um, people like Elon
1:22:58have chosen to focus on the universe and
1:23:02what's out there and the laws that
1:23:04govern the universe and that's their
1:23:06focus. I became a psychologist by
1:23:10academic training at least because my
1:23:12focus has always been on people. And if
1:23:16I had to share a closing thought to wrap
1:23:19up all of this research, I would say
1:23:21that having the most advanced AI and the
1:23:25access to the best models in the world
1:23:28won't won't determine whether your
1:23:32organization succeeds. We have to
1:23:35remember that underlying all of this
1:23:38stuff are people.
1:23:42People we as people are going through
1:23:44the most massive shift and changes that
1:23:48we ever have in our lives. And it's
1:23:52people who are feeling excited and
1:23:55exhausted and hopeful and scared often
1:24:00all at once.
1:24:02and we're all taking part in crafting
1:24:05and building this future that we hope
1:24:08still has a place for us in a few years.
1:24:13And so I keep that in mind all the time.
1:24:16Um, it's people driving these
1:24:19innovations. It's incredible people
1:24:22building these technologies. My focus
1:24:25will always be on on the people. And I
1:24:28hope that if you're out there watching
1:24:29this, you take care of yourself.
1:24:33You touch grass. You take a moment to
1:24:37reflect on what you want for yourself in
1:24:39in your career. And you remember again
1:24:42that it's fine to feel all of the
1:24:44emotions and you should just uh do
1:24:47whatever works for you to get through
1:24:49this wild period in the tech industry.
1:24:54So, kind of building on that, you you
1:24:56spend a lot of time thinking through the
1:24:59results and what people can actually do
1:25:01with this data because it's one thing to
1:25:02just report. Here's what people are
1:25:04feeling. Okay, cool. Uh what should
1:25:06people do? So, there's a really
1:25:08important section at the end of the
1:25:09report that uh that we should spend a
1:25:11little time on, which is just what where
1:25:13do we go from here? What does this tell
1:25:14us we should be doing if we can do
1:25:16anything to kind of move ourselves
1:25:17basically into a happier place? So
1:25:20starting with if you're an employee, you
1:25:23know, especially if you're an IC working
1:25:25at a company, one thing we saw people
1:25:28who feel amplified and energized
1:25:31reporting is that they actually went
1:25:34deep on specific tasks and specific jobs
1:25:38to be done rather than trying to be the
1:25:41generalist who does everything. I think
1:25:43it's the people who try to be or or lean
1:25:47too much into being a generalist who end
1:25:49up getting severely burnt out, right?
1:25:53And so I just want to caution people a
1:25:55little bit away from the narratives that
1:25:58you need to be a generalist at this
1:25:59point and just do all of the things and
1:26:02completely ignore the core parts of your
1:26:05role. I think that's a bad bad idea and
1:26:07based on what people are reporting, it's
1:26:08not going to end up with you at least
1:26:10feeling good, right? Um, I also would
1:26:14love for people to watch out for this
1:26:17squeeze being overworked for the same
1:26:20pay. We deserve better. We've linked in
1:26:24the article to a burnout test. Take the
1:26:27test. See where you're at on the burnout
1:26:30scale. And then, you know, based on
1:26:32that, reflect on where you are. Talk to
1:26:36your manager. Scope out your work again.
1:26:40you know, reccalibrate, realign with
1:26:42your manager on your work and make sure
1:26:43that you're not being uh squeezed out of
1:26:46of uh you know, being positive about
1:26:50about your role and being successful in
1:26:52your role. Um, as you're going through
1:26:54this list, let me just make sure people
1:26:56fully understand the frame there were
1:26:57because you made these two buckets
1:26:58basically things you can do based on
1:27:00this results. One is just if you're an
1:27:02employee, I think you have five pieces
1:27:04of advice. If you're a company, you have
1:27:06fiveish pieces of advice. So, just to be
1:27:08super clear, the first one is just uh
1:27:10pick a couple things you want AI to be
1:27:14useful for and just go deep on those
1:27:15couple things versus trying to spread it
1:27:17all around and AI has to be doing
1:27:19everything for you. Pick a couple things
1:27:20that you're most excited about and then
1:27:22just go really deep there. The other is
1:27:24this what you said, just watch the
1:27:25squeeze. Watch you being squeezed and
1:27:27talk to your manager about, hey, my
1:27:28scope's increased, nothing's changed
1:27:30compwise, that kind of.
1:27:31>> Absolutely. And then speak of managers.
1:27:33Again, the third point is that your
1:27:35manager clearly matters more than pretty
1:27:36much anything else when it comes to how
1:27:38you feel about work. Protect that
1:27:41relationship. Invest in that
1:27:44relationship. Build up good
1:27:46communication lines with your manager.
1:27:48[snorts]
1:27:49Learn to manage up.
1:27:52That investment in that relationship
1:27:54will pay off greatly in terms of how you
1:27:56feel. Um and then beyond that,
1:28:00um first of all, if you want to try out
1:28:03working for a smaller company or
1:28:05starting your own company, clearly there
1:28:07are some advantages to that and so it
1:28:10might be worth considering. And then if
1:28:13you're earlier in your career and you
1:28:16feel like the rungs of the ladder are
1:28:19slowly disappearing and it's becoming
1:28:21harder to climb that ladder, um I'm a
1:28:23huge believer in mentorship and the
1:28:25value of mentorship. Strong mentorship
1:28:28is still an incredibly effective thing.
1:28:31Seek the teams, seek the managers, seek
1:28:33the people who are willing to invest in
1:28:35developing you. Um it's incredibly
1:28:38valuable, especially in this era. So
1:28:41that's what we have for employees. And
1:28:44then if you're leading a team, if you're
1:28:46leading a company, then again, as we
1:28:49mentioned before, invest in your
1:28:51managers. It's probably some of the best
1:28:52money you'll ever ever spend. Um it's
1:28:55it's sad to me that only a quarter of
1:28:58our sample and perhaps tech workers in
1:29:01general rate their manager as highly
1:29:03effective. That to me is a major major
1:29:05red flag. It's a major issue in tech.
1:29:08And so if you want to improve burnout,
1:29:11if you want to elevate enjoyment, if you
1:29:14want to improve retention with these
1:29:16crazy job offers flying around from from
1:29:20AI labs, you have to invest in your
1:29:23managers. Um, and then otherwise
1:29:27manage that squeeze. People are feeling
1:29:30how AI is raising bars in ways that
1:29:33aren't sustainable. So you have to
1:29:35figure out the right level of
1:29:38expectations and the right level of
1:29:39productivity. Um don't let that bottom
1:29:44rung of the ladder rot. You know, make
1:29:47sure that you're doing what you need to
1:29:49for entry- levelvel people to advance
1:29:52because those early career people are
1:29:53also probably some of the most AI native
1:29:56people you can find, right? They find it
1:29:58very natural to use these technologies
1:30:02in some ways more so than people who
1:30:05have been uh you know too uh used to
1:30:11previous paradigms. So um and then
1:30:17selfishly [snorts] I hope that leaders
1:30:19pay a little bit more attention to
1:30:22design and research and other roles who
1:30:24are feeling uh more negative than than
1:30:28other roles. And keep in mind that this
1:30:31technology AI is lifting some people and
1:30:36destabilizing others. And you really
1:30:39have to pay attention to that. We are
1:30:41clearly not experiencing this technology
1:30:45in the same way.
1:30:47>> Amazing takeaways. I want to double down
1:30:49on that burnout survey that you um that
1:30:51you mentioned. I want to make we're
1:30:52going to link to it in the show notes,
1:30:54but I actually people like it sounds
1:30:56like dumb and why am I going to do that?
1:30:58I've like I've definitely gone through
1:30:59periods of burnout and wasn't aware that
1:31:01that was was happening and it's so
1:31:04helpful to like oh wow this is not
1:31:06normal what I'm feeling right now and
1:31:08knowing this that you're okay you're
1:31:10extremely burned out right now you you
1:31:13then you can decide what you want to do
1:31:14with that information often it's you
1:31:15know talk to a coach talk to your
1:31:16manager take some time off if you can
1:31:18but just knowing that is actually really
1:31:20powerful and he made this really great
1:31:23uh quiz based on actual science and
1:31:25research that that uh that we'll link
1:31:27you
1:31:28>> 100%.
1:31:29>> Awesome.
1:31:31Uh now, um is there anything else that
1:31:35we haven't talked about? Anything else
1:31:36you want to leave folks with? We've
1:31:38covered a lot, so I don't know if
1:31:39there's anything left. We'll obviously
1:31:41link them to the report where they can
1:31:42go deeper. We're also going to do a
1:31:44follow-up based on what comes out of
1:31:46this conversation, what we see in the
1:31:47comments because this is a lot of
1:31:49information and a lot of uh scary stuff,
1:31:51some optimistic stuff. um we can't, you
1:31:54know, we're not gonna have answers to
1:31:55everything, but we're planning to do a
1:31:56follow-up of just here's what came up.
1:31:58Let's see if we can find any answers.
1:32:00But other than that, is there anything
1:32:01else that you want to share? There's one
1:32:05interesting phenomenon that I wanted to
1:32:07to mention. um we're starting to see
1:32:10this AI guilt and in particular amongst
1:32:14people who are early in their career in
1:32:16the sense that people who are early in
1:32:18their career feel like leveraging the
1:32:21technology is a little bit like cheating
1:32:24in a sense and that guilt declines with
1:32:28seniority. Um, and in particular, we're
1:32:32seeing people in actually product
1:32:34marketing and data and analytics,
1:32:37which showed up previously in a couple
1:32:38of places, feeling the most guilt. Um,
1:32:43this is related to the well-known and
1:32:46well-established imposttor phenomenon,
1:32:49right? Like, I'm not good enough. Um,
1:32:53competent people feeling like their
1:32:54success isn't really theirs, it's
1:32:56someone else's. except this time we're
1:32:59putting that on AI on a technology
1:33:02rather than other people. Um, I would
1:33:06say to all of you who might be feeling
1:33:08that guilt, AI is an incredible
1:33:10technology. Um, leverage it. Learn how
1:33:14to leverage it. I doubt it's going away.
1:33:18Um, it's the worst it's ever going to be
1:33:20today. It's only getting better. Um, so
1:33:22there's no no reason to feel guilty
1:33:24about leveraging it. on on the contrary.
1:33:27Um I I think inevitably we all are all
1:33:30the time. So
1:33:32>> well to close note, I just want to
1:33:33appreciate you. It's incredible the work
1:33:36that you've done here and the work that
1:33:38we've done together over the many
1:33:39surveys. Uh it's just like it's like a
1:33:42side thing you do with me. Uh running
1:33:44like I don't know one of the biggest
1:33:46surveys out there around this stuff.
1:33:47Maybe the biggest analyzing it, writing
1:33:49it uh so beautifully. Uh so, uh I think
1:33:53this is a really unique and and special
1:33:55report and I appreciate you for doing
1:33:57this with me. It's uh I'm really uh I'm
1:34:00really happy about where this all goes
1:34:01and and just the what we're doing for
1:34:03folks.
1:34:04>> Thank you so so much.
1:34:05>> Well, with that, Noam, where can people
1:34:08find you online if they want to reach
1:34:09out, maybe follow-up questions, and how
1:34:11can listeners be useful to you?
1:34:13>> Yeah, so the best places to find me are
1:34:15on LinkedIn. I'd love to connect with
1:34:17anyone who wants to connect on LinkedIn
1:34:19or within Lenny's community within uh
1:34:22your Slack. Um I hang out there and
1:34:24always happy to talk to people within
1:34:26the community. So please catch me on
1:34:28LinkedIn or on Slack. That would be
1:34:30best.
1:34:30>> How can listeners be useful to you? No
1:34:32one.
1:34:32>> If you're watching this,
1:34:35what I would like to convey is that um I
1:34:37hope this research has been helpful to
1:34:39you and more than anything I hope that
1:34:41you consider participating in future
1:34:43research projects of ours. Um, I'm
1:34:46really really glad that we get to do
1:34:48these things, Lenny. And and I want to
1:34:52give all of you watching the opportunity
1:34:54to learn more about what's going on in
1:34:56tech, how you can be better at your job,
1:34:59you know, how you can feel better about
1:35:02your job and all things in in between.
1:35:05We plan to continue conducting both um
1:35:09quick polls of the community in-depth
1:35:12longer form surveys, interviews, all
1:35:16sorts of things. And so if you get a
1:35:19message, an email, a Slack about a
1:35:22research project, please participate.
1:35:24Would love to hear your voices, hear
1:35:27your perspectives, and then share that
1:35:29back with the community. And hopefully
1:35:31it makes us all better at our jobs and
1:35:34also feel better about our jobs because
1:35:36I do hope that even though much of this
1:35:39was a little bit tough and perhaps
1:35:42negative, I too Lenny am an eternal
1:35:45optimist. I'm a techno optimist. I I
1:35:48believe this is going in good directions
1:35:50and I want to make sure we capture that
1:35:52accurately. So yeah, please participate.
1:35:55We'd love to see you in our next
1:35:58research project.
1:36:00>> Excellent. Ask with that. Noam, thank
1:36:02you so much for being here. Thanks a
1:36:04bunch.
1:36:06Thank you so much for listening. If you
1:36:08found this valuable, you can subscribe
1:36:09to the show on Apple Podcasts, Spotify,
1:36:12or your favorite podcast app. Also,
1:36:14please consider giving us a rating or
1:36:16leaving a review as that really helps
1:36:18other listeners find the podcast. You
1:36:20can find all past episodes or learn more
1:36:22about the show at lennispodcast.com.
1:36:25See you in the next episode.

About this transcript

This page contains the full transcript of Why the AI’s honeymoon is ending (and tech workers are feeling it) | Noam Segal by Lenny's Podcast, generated from the public captions YouTube serves with the video. The transcript has 15,535 words across 2,281 segments, with the original timestamps preserved so you can click any line to jump to that moment in the embedded player.

What you can do with it

Use the transcript to take notes, quote the speaker, build a study guide, generate a summary with ChatGPT or Claude via the YouTube Summary tool, or export it as a timed subtitle file with YouTube to SRT. You can also re-open it in the transcriber to translate the transcript into 100+ languages.

Free YouTube transcript tool

YouTube2Text is a free YouTube transcript generator — no signup, no daily limit. Paste any YouTube link and get the full transcript instantly, with timestamps, click-to-jump, translation to 100+ languages, AI prompts for ChatGPT, Claude, and Gemini, and exports to TXT, SRT, VTT, or Markdown.