YouTube2Text

M4 MacBook Air after a month | developer's machine — Transcript

by Alex Ziskind · 3,194 words · 444 segments · language en · Watch on YouTube

Full transcript

  1. 0:00By now, you've probably already seen a
  2. 0:02bunch of M4 MacBook Air unboxings. It
  3. 0:04looks almost the same as last year. A 10
  4. 0:06core M4 chip, up to 32 GB of unified
  5. 0:10memory, wafer thin body, two lonely USBC
  6. 0:13ports, and of course, a slick new color.
  7. 0:16But the real story this year is under
  8. 0:17the hood. Proper Thunderbolt 4 lanes
  9. 0:20this time, and a surprise ability to run
  10. 0:23two external 4K screens, plus the
  11. 0:26built-in screen. I've used every single
  12. 0:28Apple Silicon Air machine ever since.
  13. 0:32Well, Apple Silicon M1. And after a
  14. 0:34month with the M4, I'm here to show you
  15. 0:36the real dev results. Docker code
  16. 0:38builds, local LLM tests, the stuff that
  17. 0:41matters to software developers. And with
  18. 0:43Apple giving us more RAM options than
  19. 0:45actual ports, which config should you
  20. 0:47get? I got some suggestions as we go
  21. 0:49through the video.
  22. 0:52The short version is that the M4 keeps
  23. 0:54boosting single core power while also
  24. 0:56speeding up multi-core tasks cuz each
  25. 0:58core is now faster. So everything's
  26. 1:00going to be faster. Geekbench single
  27. 1:02core reflects everyday snappiness,
  28. 1:03JavaScript applications in your browser.
  29. 1:05Multi-core measures how fast you can
  30. 1:08handle actual parallel jobs like Docker
  31. 1:10images or heavy code compiles. We've
  32. 1:12come a long way since M1 to M2 was about
  33. 1:15a 15% jump. M2 to M3 about 17%. And now
  34. 1:19the M4 sees around 23% gains over the
  35. 1:22M3. That's pretty good news for your dev
  36. 1:24environment. This means you can expect
  37. 1:25slightly shorter compile times, smoother
  38. 1:27multi-container docker spins, and fast
  39. 1:30hot reloads. And remember, all these
  40. 1:31MacBook Airs, the Apple silicon
  41. 1:33varieties, are fanless. The Intel one
  42. 1:36had a fan and still has a it's pretty
  43. 1:37loud, actually. You've probably seen it
  44. 1:39in my videos. Well, that means that
  45. 1:40under really heavy marathon loads, you
  46. 1:43can lose a bit of that top performance.
  47. 1:45When the machine gets really hot, it
  48. 1:47starts to throttle. And that means about
  49. 1:485 to 10% less speed than the peak
  50. 1:51benchmarks would
  51. 1:53suggest. All right, let's get practical.
  52. 1:55Real developer work is build, run,
  53. 1:57break, rebuild. And I did actually
  54. 1:59stress this M4 Air for the last month or
  55. 2:02so with every coding task that I could
  56. 2:04muster. It's actually been more than a
  57. 2:05month, but a month looks nicer in a
  58. 2:07title. I've actually gone through some
  59. 2:08of these in more detail in some of the
  60. 2:10videos, but here's the recap. Even
  61. 2:11though I'm not a C++ developer, I did
  62. 2:13some C++ tests. Quick sort on 10 million
  63. 2:16integers took about 2 minutes 34
  64. 2:17seconds, and that's 26 seconds faster
  65. 2:19than the M3. Makes sense. Single core,
  66. 2:22but full core merge sort took about 2
  67. 2:24minutes 31 seconds, almost a minute
  68. 2:26faster than the M3. Huge gains there.
  69. 2:29So, obviously, the more cores you use,
  70. 2:30the better gains you're going to see.
  71. 2:32Python matter test that I usually do.
  72. 2:34It's an interpretive bit of code that
  73. 2:35uses all the CPU cores as well. Finished
  74. 2:37in 34 seconds, beating last year's time
  75. 2:40by 15 seconds. That's crazy. And of
  76. 2:42course, speedometer 3.0 single core
  77. 2:44operations there. Chrome hit 49.3, about
  78. 2:48four points higher than the M3. So,
  79. 2:50pretty good. Now, that's not the only
  80. 2:51story here. Since this comes with 16 GB
  81. 2:54to start and 32 at the maximum, I tested
  82. 2:57those machines against each other in a
  83. 2:58video, separate video with more details.
  84. 3:00But just to give you a taste, this is
  85. 3:01where more memory really counts. I built
  86. 3:04a big NX monor repo in 3 minutes and 8
  87. 3:06seconds on the 16 GB model, but on the
  88. 3:0932 it dipped to 2 minutes and 12
  89. 3:11seconds. And I did a huge.net project
  90. 3:14compilation with 100,000 name spaces and
  91. 3:17classes. Compilation took 94 seconds on
  92. 3:19the 16 GB version and 69 seconds on the
  93. 3:2232 GB version. Pretty big difference
  94. 3:24there. Finally, Docker Compose spinning
  95. 3:27up 100 containers took about 20 seconds
  96. 3:29on the 16 GB machine and 19 on 32. Not
  97. 3:33that much of a difference in that case.
  98. 3:35Big difference on local LLMs. Pretty
  99. 3:37popular topic these days, especially on
  100. 3:40this channel. You all keep asking me
  101. 3:41about that all the time. I ran Deep Sea
  102. 3:43Coder 1.5 billion. Small model, but
  103. 3:45gives us an idea of what happens. It
  104. 3:47breezed at 55 tokens per second on the
  105. 3:4916 GB model and 63 on the
  106. 3:5332. Last year I showed how an 8 gigabyte
  107. 3:56machine runs pretty much everything.
  108. 3:58There's a few things that didn't run
  109. 4:00very well. I have a video on that, but
  110. 4:02most of the time you're going to be okay
  111. 4:03with 8 gigabytes as long as you're not
  112. 4:05running too many things at the same time
  113. 4:06obviously. But it's surprising what Mac
  114. 4:09OS can actually handle in those cases.
  115. 4:11But 16, you've got a lot more room to
  116. 4:13move around. But once you push bigger
  117. 4:15toNet builds, for example, with giant
  118. 4:17monor repos, the extra RAM shaves off
  119. 4:19big chunks of time. And of course, the
  120. 4:2132 GB machine, which I nicknamed the
  121. 4:23MacBook Air Pro. We're getting more and
  122. 4:25more Pro every year. Actually, that
  123. 4:27one's going to give you a lot more room
  124. 4:28to spread out. You can have a lot more
  125. 4:30things running at the same time and not
  126. 4:32feel the effects of things slowing down.
  127. 4:34For example, in my normal dev day is
  128. 4:36four terminal panes open running git
  129. 4:38commands and npm scripts, a couple of
  130. 4:39instances of VS Code. I have Xcode with
  131. 4:42an iOS simulator running, and
  132. 4:43occasionally I run the Android emulator,
  133. 4:46which is a beast. But this year actually
  134. 4:48they've done a good job with uh
  135. 4:50optimizing it a little bit more and it
  136. 4:52runs a little bit smoother than the
  137. 4:53previous years. Docker desktop sleeps
  138. 4:54until I really need it. But Chrome stays
  139. 4:56open with at least 25 taps. And during
  140. 4:59this whole time, the OS stays responsive
  141. 5:01even on a 16 GB machine. App switching
  142. 5:03is instant except those times when I
  143. 5:05fire up Android Studio and Docker
  144. 5:06containers all at the same time. That's
  145. 5:08when I feel a little bit of a half
  146. 5:09second hiccup. It's kind of like
  147. 5:11cramming too many people into a small
  148. 5:12car. Somebody's going to have to wait
  149. 5:14for a seat. Come on, let the old lady
  150. 5:15sit down, you rude person, you. Now, if
  151. 5:17you're expecting marathon loads or
  152. 5:19multiple virtual machines, maybe even
  153. 5:21the MacBook Air is not enough for you if
  154. 5:23you're doing that kind of stuff. In
  155. 5:24which case, you're probably going to
  156. 5:25want to dive into a MacBook Pro. And I
  157. 5:28actually tested the 32 GB MacBook Air
  158. 5:30versus the 16 GB MacBook Pro because
  159. 5:33they were the same price when they came
  160. 5:34out. Same M4 chip. The Pro, of course,
  161. 5:36has a few more features like it's a
  162. 5:39little thicker, more heat is spreading
  163. 5:41around. It has a fan, so it's gonna suck
  164. 5:43that air out or
  165. 5:45blow, whatever you want to call it. So,
  166. 5:47it's got more thermal headroom. It also
  167. 5:48has a better screen, of course, with
  168. 5:50more hertz. But here's a summary on how
  169. 5:52it did on some tests that I ran. C++
  170. 5:54quick sort single core. The Pro finished
  171. 5:56in 2 minutes 33 seconds, while the 32 GB
  172. 5:59Air took 2 minutes and 36 seconds.
  173. 6:01Obviously, they have the same cores,
  174. 6:03same chip, so very close. But maybe that
  175. 6:05fan and the Pro helped the clock speeds
  176. 6:07just a little bit. C++ merge sort all
  177. 6:10the cores all at once. 2 minutes and 18
  178. 6:12seconds on the Pro versus 2 minutes and
  179. 6:1326 seconds on the Air. Here we saw the
  180. 6:16Pro's cooling actually give an edge to a
  181. 6:18sustained load. That huge net
  182. 6:19compilation with 100,000 classes, 69
  183. 6:22seconds on the 32 GB Air and 87 seconds
  184. 6:25on the 16 GB Pro. Wow, that extra RAM
  185. 6:28saves you from swapping. So, the Air
  186. 6:30crushed that test. The Air beat the Pro.
  187. 6:32The Manroad Python test, 29 seconds on
  188. 6:35the Pro, 34 on the Air. Again, this is
  189. 6:37using all the cores and it's pretty
  190. 6:39intense. So, I think the fan helped out
  191. 6:42here again. So, if your job is mostly
  192. 6:43huge CPU sprints, the Pro's fans keeps
  193. 6:46higher speeds under load. If you need
  194. 6:48more than 16 GB of memory and do mostly
  195. 6:50short builds, the 32 GB Air can outrun a
  196. 6:53Pro on RAM heavy tasks and it stays
  197. 6:55silent. Okay, real talk. Dropping 2
  198. 6:58grand on a MacBook up front, that's a
  199. 7:00whole car repair or a vacation or half a
  200. 7:03Thunderbolt dock. So, here's a smarter
  201. 7:05move. upgraded. They let you pick the
  202. 7:07exact MacBook you want. Custom specs,
  203. 7:10more RAM, extra storage, and instead of
  204. 7:12paying everything up front, you spread
  205. 7:14it out over 36 months. Interest starts
  206. 7:17from 0% depending on your eligibility.
  207. 7:19Even better, after 24 months, you just
  208. 7:21send the MacBook back in the box that
  209. 7:23they give you, and boom, new MacBook
  210. 7:26time. No sketchy marketplace listings.
  211. 7:29No, is this still available messages at
  212. 7:312 a.m. They work with Apple Premiere
  213. 7:33Partners. They have a 4.8 in a trust
  214. 7:35pilot rating and yeah, even folks from
  215. 7:38OpenAI back then. The best part is you
  216. 7:40can keep the cash you didn't drop up
  217. 7:42front earning 4% interest in a savings
  218. 7:44account. So go to getupgraded.com link
  219. 7:47in the description. Use code alex40 and
  220. 7:49get 40 bucks off your next MacBook.
  221. 7:51That's
  222. 7:52getupgraded.com. Code
  223. 7:56alex40. There was a few days when I went
  224. 7:58on a trip and I spent full work days
  225. 8:00relying only on the 16 GB M4 Air. Even
  226. 8:03with VS Code, Xcode, Chrome, Notion,
  227. 8:06Docker, Compose running, I saw no real
  228. 8:08lag. At worst, Mac OS swapped out about
  229. 8:101 GB. That's when it's using the SSD
  230. 8:12instead of the RAM when everything
  231. 8:14doesn't fit into RAM. Not a good thing
  232. 8:15to have happen. So, keep an eye on that
  233. 8:17if you have a low memory machine and
  234. 8:19you're using a lot of applications at
  235. 8:20the same time. But nonetheless, Windows
  236. 8:22switching felt smooth. Programs pop back
  237. 8:24in pretty easily. Now, I did notice that
  238. 8:26after 10-minute full core compile, the
  239. 8:28palm rest warmed up to around 46°. Not
  240. 8:31painful, but definitely toasty compared
  241. 8:33to older models. In fact, coming back to
  242. 8:35the M4 Air and putting your palms on it,
  243. 8:37it's always a little bit surprising
  244. 8:39because the palm rests are warm. Whereas
  245. 8:41in the previous models, they're kind of
  246. 8:43cool. So, you expect that. So, that's
  247. 8:45something new. Not great, but it's new.
  248. 8:47As far as battery goes, it seems like I
  249. 8:49could leave this thing on for days and
  250. 8:51it'll just keep working and not running
  251. 8:52out of battery. But, I wanted to be a
  252. 8:54little bit more scientific about it. So,
  253. 8:57I said a little bit more scientific, not
  254. 8:58a lot. Okay, just calm down. calm down.
  255. 9:00I did my developer loop, which I call it
  256. 9:03now. It's a dev test that I run that
  257. 9:05simulates what a dev would do, what I
  258. 9:07would do actually. Not all devs do the
  259. 9:09same stuff, but it has a bunch of
  260. 9:11programs open like things that I use,
  261. 9:13notion. It's got a little bit of music
  262. 9:15playing. The screen is set to full
  263. 9:16brightness, to-do list, uh 15 browser
  264. 9:19tabs. I have a couple of instances of
  265. 9:21terminal open and VS code and it runs
  266. 9:23through a loop. A half an hour loop of
  267. 9:25things that we do like coding, building
  268. 9:27code, running interpreted code.
  269. 9:29Basically, things that I covered
  270. 9:30already, but in a loop. And of course,
  271. 9:32watching videos is one of the things
  272. 9:33that we do. I know you do that. I just
  273. 9:36know it. I I have a feeling. How do I
  274. 9:37know this? I don't know. You tell me. I
  275. 9:39actually ran this one twice and it got
  276. 9:41about the same results on both the uh 32
  277. 9:44GB version and the 16 GB version. Now,
  278. 9:47the M4 MacBook Air actually has a
  279. 9:49slightly bigger battery,
  280. 9:5153.8 watt hours as opposed to 52.6 from
  281. 9:55last year, the M3. And look at these
  282. 9:57two. The M3 MacBook Air and the M4
  283. 9:59MacBook Air are very close to the number
  284. 10:03of minutes that they lasted under my
  285. 10:05test. About 330 minutes for the M3. The
  286. 10:08M4 lasted just a little bit less. 315
  287. 10:11maybe. If you interpolate that line
  288. 10:13down, it's going to be like 328 or
  289. 10:16something like that, but very close. How
  290. 10:18does this compare with some of these
  291. 10:19other machines that I tested? There's
  292. 10:20the Galaxy Book with the X Elite. The
  293. 10:23MacBook Pro 14, the M4 version lasted a
  294. 10:26little bit longer than both of those.
  295. 10:28It's got a slightly bigger battery.
  296. 10:29Here's the M4 Pro lasted quite a bit
  297. 10:32longer. The M2 Max is over here, a
  298. 10:352-year-old machine that I've been daily
  299. 10:37driving. So, yeah, the battery is going
  300. 10:39to go down a little bit. And here's the
  301. 10:40MacBook Pro 16 M4 Max. That's my current
  302. 10:43daily driver. Just about the same as the
  303. 10:45M3 MacBook Air. If there's anything that
  304. 10:47lasted a significant amount longer,
  305. 10:49here's a Surface Laptop 7 with the X
  306. 10:52Elite chip inside of it. That one lasted
  307. 10:54the longest. But we're all in a
  308. 10:56territory now where we don't really need
  309. 10:58to care about charging our laptop
  310. 11:00throughout the day. If you're working
  311. 11:02normal days and you're not Elon Musk who
  312. 11:04works 18our days, then you may be okay.
  313. 11:06That's what he said. Anyway, I I don't
  314. 11:08know. Maybe that he said that. Don't
  315. 11:09quote me on that. So, let's talk about
  316. 11:11who benefits from this
  317. 11:14upgrade. The M4 Air is still a paper
  318. 11:17thin laptop that slips into a backpack
  319. 11:19easily. With Apple's latest chip, it can
  320. 11:21handle serious deaf tasks on the go, no
  321. 11:24problem. So, for students and general
  322. 11:26developers, if you're in class or
  323. 11:27meetings all day, coding, note-taking,
  324. 11:29or Zoom calling, you last 8 to 10 hours
  325. 11:31on battery, no problem. You also won't
  326. 11:33distract the room with fan noise. A
  327. 11:36second benefit, people that are
  328. 11:37traveling, this is really nice. I've
  329. 11:39been really enjoying carrying something
  330. 11:41like this as opposed to the large
  331. 11:43MacBook Pro, especially when I'm walking
  332. 11:46around with my backpack. It really makes
  333. 11:47a big difference. It also will fit on
  334. 11:49cramped airplane trays. I brought a
  335. 11:52prop. This is the size of a typical
  336. 11:55airplane tray. Here's the MacBook Pro
  337. 11:5716in. Where do you put your mouse? Where
  338. 11:58do you put your food? Where do you put
  339. 12:00your coffee? And here is the MacBook
  340. 12:02Air. M4, M2, M3. Same size. you have a
  341. 12:05little bit more room to work with. Candy
  342. 12:07bar goes here. I don't know, your
  343. 12:08chicken nuggets go here. You can also
  344. 12:10charge it with a smaller USBC adapter, a
  345. 12:13power brick that perhaps you're
  346. 12:14borrowing because you forgot
  347. 12:16[Music]
  348. 12:17yours. If you crave more ports, built-in
  349. 12:20HDMI, or need that 120 Hz display, the
  350. 12:23MacBook Pro is the one to go with, of
  351. 12:25course. But I'm going to tell you this.
  352. 12:26If you're used to 60
  353. 12:28Hz, don't spoil yourself. Don't try the
  354. 12:31120 Hz MacBook Pro display. It's going
  355. 12:33to ruin you. It's going to spoil your
  356. 12:35day cuz you're going to want it.
  357. 12:37However, after working on the MacBook
  358. 12:40Air for that trip that I did for a
  359. 12:42while, you kind of get used to 60 Hz.
  360. 12:44Again, there is a little bit of a
  361. 12:45difference. You notice it, but it's
  362. 12:47something you get used to, especially if
  363. 12:49all you're doing is just looking at
  364. 12:50code. That's not all we're doing, is it?
  365. 12:52We're watching YouTube, too. Also, the
  366. 12:54MacBook Pro, the fan gives you better
  367. 12:56sustained speeds. So, if you got large
  368. 12:58compilations to do, it's going to be a
  369. 13:00little faster. If your tasks are small
  370. 13:02and fit inside two-minute bursts, the
  371. 13:04Air can match or beat the Pro,
  372. 13:05especially with that 32 GB version. So,
  373. 13:08if you are going with the Air, here are
  374. 13:09some recommended builds for everyday
  375. 13:11work. 16 GB of RAM, 1 TB SSD, that's
  376. 13:14fine. Apple charges a lot for those
  377. 13:17upgrades of the storage, and it just h
  378. 13:19you can get something like this easy.
  379. 13:22It's another thing to carry around, I
  380. 13:24know, but if you want to just throw it
  381. 13:25in your backpack, and if you already
  382. 13:26have a backpack, then it's not such a
  383. 13:28big deal to carry extra stuff with you.
  384. 13:30But data only, not programs of course.
  385. 13:32So if you have a lot of programs like
  386. 13:35Xcode is a honker, right? Android Studio
  387. 13:37is big, Docker, all those programs are
  388. 13:39going to add up. 256 GB that I have on
  389. 13:43the base model here, it's a headache.
  390. 13:44After Xcode, Android SDKs, Docker
  391. 13:47images, node modules, you're already
  392. 13:49almost out of space after installing all
  393. 13:51that stuff. Forget virtual machines. 512
  394. 13:54GB works if you move large assets to an
  395. 13:56external drive, but you'll also be
  396. 13:58worried about constant space management.
  397. 14:001 TB right now is pretty good. Gives you
  398. 14:01enough room to breathe. You can stash
  399. 14:03multiple virtual machines plus local git
  400. 14:06clones, maybe a couple of small LLMs.
  401. 14:08For heavy multitasking, 32 gigs of RAM
  402. 14:11is a really good option this year,
  403. 14:13especially on a MacBook Air. It's pretty
  404. 14:14amazing that Apple added that option.
  405. 14:16There's also the 24 GB option, and I'm
  406. 14:19not too sure about that one. I don't
  407. 14:20know. Yeah, it's there, but if you need
  408. 14:22RAM, get 32. If you don't need that much
  409. 14:25RAM, go for 16. 24 is like it's like you
  410. 14:28can't decide. But if you get more RAM,
  411. 14:29you're not going to be sorry. If you get
  412. 14:31less RAM, you might be sorry. So, the M4
  413. 14:34Air keeps its single core supremacy. If
  414. 14:36your compiles stay under 2 minutes, it
  415. 14:38actually rivals a 16 GB MacBook Pro. It
  416. 14:40runs warm enough to notice, but not too
  417. 14:42hot enough to harm performance. It will
  418. 14:44throttle if you go push it too hard. But
  419. 14:46also, in the winter time, you can think
  420. 14:47of it as a built-in hand warmer. The
  421. 14:50thing has silence and stamina. Eight
  422. 14:52coding hours, still near half the
  423. 14:54battery, and no fan noise is a beautiful
  424. 14:56thing to see at the end of the day. So,
  425. 14:58the sweet spot is 16 GB of RAM, 1 TB of
  426. 15:01storage. Go 32 if you can. You won't be
  427. 15:04sorry if you do, and avoid the 256 GB
  428. 15:07storage tier if you value your sanity.
  429. 15:10If that profile matches your workflow,
  430. 15:11the M4 Air might be the most portable
  431. 15:14dev laptop you can buy. Also, don't
  432. 15:16forget if you do need more ports, you
  433. 15:17can always get a little gadget like
  434. 15:20this. I have a couple of other docks.
  435. 15:22This one is easy cuz it doesn't have a
  436. 15:24power plug. It's not Thunderbolt, but it
  437. 15:26just gives you more stuff, even an HDMI
  438. 15:29port. I'll leave a link to that one down
  439. 15:30below. It's pretty nice. I think that's
  440. 15:32going to wrap it up, folks. Thanks for
  441. 15:33watching, and I will see you in the next
  442. 15:35video.
  443. 15:37[Applause]
  444. 15:38[Music]

About this transcript

This page contains the full transcript of M4 MacBook Air after a month | developer's machine by Alex Ziskind, generated from the public captions YouTube serves with the video. The transcript has 3,194 words across 444 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 choose a caption track available on the video.

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, caption language selection, AI prompts for ChatGPT, Claude, and Gemini, and exports to TXT, SRT, VTT, or Markdown.