{"video_id":"FgG57u78ZcI","title":"How can I justify spending $10K on a Mac  - Mac Studio M3 Ultra","channel":"ShortCircuit","show":"ShortCircuit","published_at":"2025-05-04T14:58:16Z","duration_s":696,"segments":[{"start_s":0.16,"end_s":6.96,"text":"You guys just saw our income breakdown, so you know that unless you buy an awful","speaker":null,"is_sponsor":0},{"start_s":4.319,"end_s":11.92,"text":"lot of Framework hoodies, I am not making back what I spent on this brand","speaker":null,"is_sponsor":1},{"start_s":9.2,"end_s":17.359,"text":"new M3 Ultra Max Studio. So, the question is, how can we possibly justify","speaker":null,"is_sponsor":0},{"start_s":14.96,"end_s":22.96,"text":"buying it? And the answer may surprise you. You see, depending on what you're","speaker":null,"is_sponsor":0},{"start_s":19.68,"end_s":26.72,"text":"using it for, the M3 Ultra Studio is","speaker":null,"is_sponsor":0},{"start_s":22.96,"end_s":28.24,"text":"actually, dare I say it, a good deal.","speaker":null,"is_sponsor":0},{"start_s":26.72,"end_s":33.6,"text":"Not for everyone, of course. If all you're doing is some light photo editing, you're going to be better off","speaker":null,"is_sponsor":0},{"start_s":30.64,"end_s":37.6,"text":"doing that on your phone. But if you work with AI, for instance, or if you","speaker":null,"is_sponsor":0},{"start_s":36.0,"end_s":42.719,"text":"have heavy demands when it comes to content creation, then even with our","speaker":null,"is_sponsor":0},{"start_s":40.559,"end_s":48.399,"text":"configuration, which has the topsp spec GPU and cost us about","speaker":null,"is_sponsor":0},{"start_s":46.039,"end_s":54.399,"text":"$10,000, but this little puppy could end up making you a lot more than you spent.","speaker":null,"is_sponsor":0},{"start_s":52.079,"end_s":58.719,"text":"Let's take a look after, of course, we uh look at the accessories. There's a","speaker":null,"is_sponsor":0},{"start_s":56.48,"end_s":58.719,"text":"power","speaker":null,"is_sponsor":0},{"start_s":60.76,"end_s":69.36,"text":"cable and paper. And that's it. So, after that","speaker":null,"is_sponsor":0},{"start_s":67.76,"end_s":75.36,"text":"very short side quest, let's take a closer look at the machine itself. Starting at the front, we've got an SDXC","speaker":null,"is_sponsor":0},{"start_s":72.799,"end_s":79.36,"text":"reader as well as two Thunderbolt ports. And these are one of the big","speaker":null,"is_sponsor":0},{"start_s":77.2,"end_s":83.36,"text":"differentiators for the new Mac Studio. Whether we're talking about the topsp","speaker":null,"is_sponsor":0},{"start_s":80.96,"end_s":90.0,"text":"spec M3 Ultra version that we've got here or the lesser","speaker":null,"is_sponsor":0},{"start_s":86.439,"end_s":92.96,"text":"M4 Max version that has faster cores,","speaker":null,"is_sponsor":0},{"start_s":90.0,"end_s":96.56,"text":"but not nearly as many of them. And the reason that these are such a big deal is","speaker":null,"is_sponsor":0},{"start_s":94.479,"end_s":101.439,"text":"because now they are Thunderbolt 5, which even under normal circumstances","speaker":null,"is_sponsor":0},{"start_s":99.2,"end_s":107.439,"text":"offers up to double the performance of last gen Thunderbolt 4 because it","speaker":null,"is_sponsor":0},{"start_s":103.759,"end_s":109.84,"text":"transitions from PCIe Gen 3 to PCIe Gen","speaker":null,"is_sponsor":0},{"start_s":107.439,"end_s":114.2,"text":"4. That means more bandwidth for things like connected SSDs or card readers or","speaker":null,"is_sponsor":0},{"start_s":112.88,"end_s":120.399,"text":"external GPUs, at least on non-MAC machines. But","speaker":null,"is_sponsor":0},{"start_s":118.0,"end_s":125.68,"text":"it also has another trick up its sleeve. Instead of operating at 80 Gbit 80 Gbit","speaker":null,"is_sponsor":0},{"start_s":123.68,"end_s":132.4,"text":"transmission and receive, it can actually stack it and do up to 120 GB","speaker":null,"is_sponsor":0},{"start_s":129.119,"end_s":134.8,"text":"transmission and 40 GB receive in boost","speaker":null,"is_sponsor":0},{"start_s":132.4,"end_s":140.239,"text":"mode. This allows the new Mac Studio with its Thunderbolt 5 ports to drive","speaker":null,"is_sponsor":0},{"start_s":137.52,"end_s":145.76,"text":"extremely highresolution, high refresh rate displays, and a lot of them. Apple","speaker":null,"is_sponsor":0},{"start_s":143.36,"end_s":149.36,"text":"boasts it'll do up to eight connected displays in the ultra configuration.","speaker":null,"is_sponsor":0},{"start_s":148.0,"end_s":154.8,"text":"Now, let's take a look at the back where we find, that's right, another four of","speaker":null,"is_sponsor":0},{"start_s":153.12,"end_s":158.239,"text":"those Thunderbolt 5 ports, allowing you to connect pretty much whatever you","speaker":null,"is_sponsor":0},{"start_s":156.4,"end_s":161.76,"text":"could possibly want. remembering, of course, that Thunderbolt can daisy","speaker":null,"is_sponsor":0},{"start_s":159.68,"end_s":166.879,"text":"chain, so you can have up to six devices off of each of those six ports. We also","speaker":null,"is_sponsor":0},{"start_s":164.56,"end_s":171.92,"text":"get a 10 gig Ethernet connection, power input, and Mickey Mouse style, two USBA","speaker":null,"is_sponsor":0},{"start_s":169.84,"end_s":175.599,"text":"ports that operate at up to 5 GB per second, an HDMI port, as well as a","speaker":null,"is_sponsor":0},{"start_s":174.0,"end_s":180.959,"text":"headphone jack, and of course, conveniently located rear power button.","speaker":null,"is_sponsor":0},{"start_s":178.64,"end_s":184.64,"text":"We also get a nice beefy looking exhaust to go along with our nice beefy intake","speaker":null,"is_sponsor":0},{"start_s":183.12,"end_s":188.879,"text":"on the bottom. I haven't heard it in person, though I will very shortly. The","speaker":null,"is_sponsor":0},{"start_s":186.239,"end_s":193.76,"text":"Labs tells me is pretty darn quiet. A Kensington lock for the bottom. And","speaker":null,"is_sponsor":0},{"start_s":191.36,"end_s":198.72,"text":"that's about it. It's just a big block of aluminum. A big block of aluminum","speaker":null,"is_sponsor":0},{"start_s":196.239,"end_s":203.04,"text":"full of some very impressive specs. Let's talk speeds and feeds while I get","speaker":null,"is_sponsor":0},{"start_s":200.56,"end_s":208.48,"text":"this thing set up. The M3 Ultra SoC in here is equipped with up to 32 CPU","speaker":null,"is_sponsor":0},{"start_s":205.68,"end_s":213.36,"text":"cores. That is eight more than Apple's last gen flagship. But perhaps more","speaker":null,"is_sponsor":0},{"start_s":211.04,"end_s":217.44,"text":"importantly than that, all eight of the new cores are performance rather than","speaker":null,"is_sponsor":0},{"start_s":215.92,"end_s":222.959,"text":"efficiency cores, meaning this thing should be a multi-threaded beast. In","speaker":null,"is_sponsor":0},{"start_s":220.239,"end_s":227.2,"text":"addition to that, we get a new 32 core neural engine that is apparently much","speaker":null,"is_sponsor":0},{"start_s":224.879,"end_s":232.879,"text":"faster than last gen. And on top of that, we get up to an 80 core GPU with","speaker":null,"is_sponsor":0},{"start_s":229.92,"end_s":237.28,"text":"up to 800 gigabytes per second of combined memory bandwidth that is shared","speaker":null,"is_sponsor":0},{"start_s":235.04,"end_s":242.68,"text":"across all of those processors. This was a really key Apple innovation in the","speaker":null,"is_sponsor":0},{"start_s":239.04,"end_s":245.92,"text":"last few years, their unified memory. Of","speaker":null,"is_sponsor":0},{"start_s":242.68,"end_s":249.04,"text":"course, fast memory is great, but when","speaker":null,"is_sponsor":0},{"start_s":245.92,"end_s":251.68,"text":"it comes to AI, more memory is in some","speaker":null,"is_sponsor":0},{"start_s":249.04,"end_s":255.519,"text":"cases more important, and that is where this thing is really appealing to AI","speaker":null,"is_sponsor":0},{"start_s":253.68,"end_s":262.32,"text":"developers. It can be equipped with up to 512 GB of LPDDR5 memory. That means","speaker":null,"is_sponsor":0},{"start_s":260.4,"end_s":269.04,"text":"that compared to last gen, which topped out at 192, we can run much larger AI","speaker":null,"is_sponsor":0},{"start_s":266.8,"end_s":273.52,"text":"models. And that ignores the fact that the hardware is faster, allowing us to","speaker":null,"is_sponsor":0},{"start_s":271.28,"end_s":277.44,"text":"run them in a more performant manner. Of course, it's not all for AI. Apple","speaker":null,"is_sponsor":0},{"start_s":275.52,"end_s":282.56,"text":"loaded the M3 Ultra up with a new media engine that they claim supports up to 24","speaker":null,"is_sponsor":0},{"start_s":280.08,"end_s":287.44,"text":"streams of ProRes playback in addition to supporting hardware encode and decode","speaker":null,"is_sponsor":0},{"start_s":285.12,"end_s":293.919,"text":"of many of the most popular formats, including AV1 decoding. Note that I said","speaker":null,"is_sponsor":0},{"start_s":290.4,"end_s":296.639,"text":"AV1 decoding, though. One major omission","speaker":null,"is_sponsor":0},{"start_s":293.919,"end_s":302.0,"text":"is that Apple's SOC's still do not support AV1 encoding. And this is true","speaker":null,"is_sponsor":0},{"start_s":299.04,"end_s":306.56,"text":"of both the M3 Ultra configuration as well as the M4 Max configuration. Either","speaker":null,"is_sponsor":0},{"start_s":305.199,"end_s":311.919,"text":"way, these machines are going to be Photoshop and Premier Powerhouses.","speaker":null,"is_sponsor":0},{"start_s":308.88,"end_s":315.759,"text":"Because even limited to CPUAV1 encoding,","speaker":null,"is_sponsor":0},{"start_s":311.919,"end_s":317.44,"text":"the M3 Ultra beat a Threadripper 7980X,","speaker":null,"is_sponsor":0},{"start_s":315.759,"end_s":322.72,"text":"a chip that consumes more power than this entire box and has twice as many","speaker":null,"is_sponsor":0},{"start_s":319.639,"end_s":324.08,"text":"cores. Pretty impressive. It's just","speaker":null,"is_sponsor":0},{"start_s":322.72,"end_s":328.24,"text":"worth noting that it doesn't support hardware encode, which is a shame when","speaker":null,"is_sponsor":0},{"start_s":325.759,"end_s":332.479,"text":"you look at the incredible performance of Apple's hardware encoder for H.264.","speaker":null,"is_sponsor":0},{"start_s":331.12,"end_s":337.039,"text":"Now, another thing that's worth noting is that Apple doesn't have a clear","speaker":null,"is_sponsor":0},{"start_s":334.96,"end_s":342.56,"text":"flagship with the Mac Studio anymore. I mean, they obviously have one that costs","speaker":null,"is_sponsor":0},{"start_s":338.56,"end_s":344.96,"text":"more, the M3 Ultra version, but the M4","speaker":null,"is_sponsor":0},{"start_s":342.56,"end_s":349.28,"text":"Max version may end up outperforming it in some cases. Now, we don't have an M4","speaker":null,"is_sponsor":0},{"start_s":346.8,"end_s":353.919,"text":"Max yet, but we do have an M4 Pro machine. And we found that in cases","speaker":null,"is_sponsor":0},{"start_s":351.28,"end_s":359.52,"text":"where singlethreaded CPU performance is more important than having a big beefy","speaker":null,"is_sponsor":0},{"start_s":356.08,"end_s":361.84,"text":"GPU or lots of CPU cores, it could end","speaker":null,"is_sponsor":0},{"start_s":359.52,"end_s":366.08,"text":"up being the more performant option. Of course, you'll also be limited to a","speaker":null,"is_sponsor":0},{"start_s":363.199,"end_s":370.56,"text":"maximum of 64 GB of memory. Now, before we go any further into performance,","speaker":null,"is_sponsor":0},{"start_s":367.84,"end_s":374.759,"text":"let's thank our sponsor, Dbrand. Uh, Dbrand told me they want you to go","speaker":null,"is_sponsor":1},{"start_s":371.759,"end_s":378.759,"text":"outside and touch grass.","speaker":null,"is_sponsor":1},{"start_s":374.759,"end_s":381.68,"text":"Okay, that shouldn't be that hard. I see","speaker":null,"is_sponsor":1},{"start_s":378.759,"end_s":386.72,"text":"some over there. I think there's some grass in there. There you go.","speaker":null,"is_sponsor":1},{"start_s":384.479,"end_s":391.199,"text":"Oh, sorry, boss man. I misread. They actually want you to touch grass on your","speaker":null,"is_sponsor":1},{"start_s":388.479,"end_s":396.72,"text":"Mac Studio. Oh, this. They actually told me they","speaker":null,"is_sponsor":1},{"start_s":394.4,"end_s":400.8,"text":"were working on this a while back. I didn't know they were ready to ship yet.","speaker":null,"is_sponsor":1},{"start_s":398.479,"end_s":407.6,"text":"That is really impressive. Apparently, every square centimeter contains a,000","speaker":null,"is_sponsor":1},{"start_s":404.96,"end_s":411.759,"text":"grass strands. Man, you can really feel the density of it. It's like having a","speaker":null,"is_sponsor":1},{"start_s":410.0,"end_s":416.479,"text":"putting green on the top of your Mac Studio. This is not an April Fool's","speaker":null,"is_sponsor":1},{"start_s":414.319,"end_s":419.68,"text":"joke. This is a real product. Well, like the Touch Grass skin is a real product.","speaker":null,"is_sponsor":1},{"start_s":418.4,"end_s":426.319,"text":"They don't make skins for the Mac Studio, so get wrecked. I guess Dbrand has","speaker":null,"is_sponsor":1},{"start_s":424.4,"end_s":429.599,"text":"fantastic guides on their site for how to apply their skins, but I don't give a","speaker":null,"is_sponsor":1},{"start_s":428.319,"end_s":433.759,"text":"if those guys want me to follow their guide, they're going to have to pay me more and be a little","speaker":null,"is_sponsor":1},{"start_s":437.319,"end_s":443.52,"text":"nicer. There you have it, folks. This is as real of grass that you can stick to","speaker":null,"is_sponsor":1},{"start_s":441.599,"end_s":448.039,"text":"your phone. It's ultra limited edition. Once it's all sold out, it's sold out.","speaker":null,"is_sponsor":1},{"start_s":445.44,"end_s":450.96,"text":"You can get it now at shortlininus.com. All right, let's talk","speaker":null,"is_sponsor":1},{"start_s":449.599,"end_s":454.8,"text":"a little bit more about the Mac Studio, shall we? I want to fire up some kind of","speaker":null,"is_sponsor":0},{"start_s":453.039,"end_s":458.88,"text":"a load and see how loud this thing gets for myself. Let's go with Blender. We're","speaker":null,"is_sponsor":0},{"start_s":456.72,"end_s":462.0,"text":"going to do the Barberhop render. I'm going to see if I can hear this","speaker":null,"is_sponsor":0},{"start_s":464.36,"end_s":472.479,"text":"thing. I mean, surely it's got to ramp up. This is CPU mode, too. This is with","speaker":null,"is_sponsor":0},{"start_s":469.759,"end_s":476.24,"text":"all the CPU cores going full tilt. I can barely hear it. And I'm pretty much the","speaker":null,"is_sponsor":0},{"start_s":474.56,"end_s":479.84,"text":"biggest diva there is when it comes to computer noise. Like to the point where","speaker":null,"is_sponsor":0},{"start_s":477.68,"end_s":484.639,"text":"I put my computer in a separate room, so I don't have to hear it. If I could game","speaker":null,"is_sponsor":0},{"start_s":482.72,"end_s":492.08,"text":"on a Mac Studio, I probably wouldn't even bother. To be clear, I can game on","speaker":null,"is_sponsor":0},{"start_s":488.56,"end_s":493.759,"text":"a Mac Studio. Uh the performance is fine","speaker":null,"is_sponsor":0},{"start_s":492.08,"end_s":497.919,"text":"whether we're looking at games that rely on Rosetta or whether we're looking at","speaker":null,"is_sponsor":0},{"start_s":496.0,"end_s":502.879,"text":"games that run natively on Apple Silicon.","speaker":null,"is_sponsor":0},{"start_s":499.759,"end_s":505.84,"text":"But what it isn't is a tremendous value","speaker":null,"is_sponsor":0},{"start_s":502.879,"end_s":510.72,"text":"and the compatibility is still a major struggle for the macOS platform","speaker":null,"is_sponsor":0},{"start_s":507.68,"end_s":513.12,"text":"compared to Windows or even Linux these","speaker":null,"is_sponsor":0},{"start_s":510.72,"end_s":517.039,"text":"days. With that said, I still feel that GPU benchmarks are relevant because they","speaker":null,"is_sponsor":0},{"start_s":515.36,"end_s":521.36,"text":"give us some idea of what Apple has improved from generation to generation.","speaker":null,"is_sponsor":0},{"start_s":519.279,"end_s":525.68,"text":"And at least when it comes to gaming, it looks like not much. Most of this","speaker":null,"is_sponsor":0},{"start_s":524.24,"end_s":531.56,"text":"performance difference compared to the M2 Ultra could be explained by going","speaker":null,"is_sponsor":0},{"start_s":527.68,"end_s":534.0,"text":"from 76 to 80 GPU cores in the maximum","speaker":null,"is_sponsor":0},{"start_s":531.56,"end_s":538.64,"text":"configuration. But what's curious is that if I wasn't running this in CPU","speaker":null,"is_sponsor":0},{"start_s":536.399,"end_s":546.36,"text":"mode, if I was running Blender in GPU mode, it is much faster in GPU Blender","speaker":null,"is_sponsor":0},{"start_s":543.32,"end_s":549.76,"text":"rendering. And the same holds true in","speaker":null,"is_sponsor":0},{"start_s":546.36,"end_s":552.399,"text":"AI. Actually, in AI, performance is up","speaker":null,"is_sponsor":0},{"start_s":549.76,"end_s":558.32,"text":"across the board. So depending on the application, it'll use either the neural","speaker":null,"is_sponsor":0},{"start_s":555.12,"end_s":560.8,"text":"processor, the GPU, or the CPU,","speaker":null,"is_sponsor":0},{"start_s":558.32,"end_s":566.399,"text":"typically in that order in order to get the best performance out of the M3 Ultra","speaker":null,"is_sponsor":0},{"start_s":563.04,"end_s":569.92,"text":"SOC. And we found that across the board,","speaker":null,"is_sponsor":0},{"start_s":566.399,"end_s":572.56,"text":"the M3 Ultra is much faster than the M2","speaker":null,"is_sponsor":0},{"start_s":569.92,"end_s":576.64,"text":"Ultra in AI. And that's in addition to being able to run much larger models.","speaker":null,"is_sponsor":0},{"start_s":574.16,"end_s":581.6,"text":"So, this was the main justification that Nick, and I'm naming him, Nick from the","speaker":null,"is_sponsor":0},{"start_s":579.04,"end_s":588.88,"text":"lab gave me for us buying this machine because we actually use our 192 GB M2","speaker":null,"is_sponsor":0},{"start_s":585.36,"end_s":590.56,"text":"Ultra Studio daily to run local LLMs","speaker":null,"is_sponsor":0},{"start_s":588.88,"end_s":596.32,"text":"that help with programming for our coders. He says, and he promises me that","speaker":null,"is_sponsor":0},{"start_s":594.08,"end_s":602.64,"text":"this machine will justify itself because we'll be able to run better LLMs. Is","speaker":null,"is_sponsor":0},{"start_s":598.56,"end_s":603.92,"text":"that right, Nick? Yep. Okay. Oh, wow.","speaker":null,"is_sponsor":0},{"start_s":602.64,"end_s":608.48,"text":"I'm just looking a little closer at those gaming results. This thinging","speaker":null,"is_sponsor":0},{"start_s":605.92,"end_s":613.36,"text":"sucks for gaming. This is like on par with a 4060 Ti. The way it kind of keeps","speaker":null,"is_sponsor":0},{"start_s":611.2,"end_s":619.68,"text":"up in Blender, though, is so bizarre to me. Like, yeah, it's slower than a 7980X","speaker":null,"is_sponsor":0},{"start_s":616.56,"end_s":623.04,"text":"with an RTX 5090, but like not by that","speaker":null,"is_sponsor":0},{"start_s":619.68,"end_s":625.12,"text":"much, you know? Like a 4060 Ti would get","speaker":null,"is_sponsor":0},{"start_s":623.04,"end_s":630.88,"text":"crushed by that configuration, and this doesn't. What a strange strange machine.","speaker":null,"is_sponsor":0},{"start_s":628.56,"end_s":638.399,"text":"And boy, will you ever pay for it. The M4 Max configuration starts at $2,000.","speaker":null,"is_sponsor":0},{"start_s":634.64,"end_s":641.44,"text":"And the M3 Ultra starts at $4,000, going","speaker":null,"is_sponsor":0},{"start_s":638.399,"end_s":644.32,"text":"up as high as about $14,000 if you max","speaker":null,"is_sponsor":0},{"start_s":641.44,"end_s":649.12,"text":"out both the RAM like we did, and the storage, which we did not. A terabyte is","speaker":null,"is_sponsor":0},{"start_s":647.839,"end_s":653.6,"text":"fine when you're going to be accessing most of your storage over the network.","speaker":null,"is_sponsor":0},{"start_s":651.68,"end_s":658.24,"text":"Oh, with that said, could we maybe try to do a storage upgrade for an LT DIY? I","speaker":null,"is_sponsor":0},{"start_s":657.2,"end_s":664.16,"text":"don't know if there's a way to do that on this one yet. We'll have to figure that out because of course like other","speaker":null,"is_sponsor":0},{"start_s":661.839,"end_s":668.72,"text":"Apple silicon machines, Apple does not officially sanction any kind of storage","speaker":null,"is_sponsor":0},{"start_s":666.8,"end_s":672.8,"text":"or memory upgrade after the fact. You got to buy it and that's what you get.","speaker":null,"is_sponsor":0},{"start_s":671.04,"end_s":679.04,"text":"And actually, for that matter, I don't think there's any official way to upgrade like the the Wi-Fi chip in it","speaker":null,"is_sponsor":0},{"start_s":675.839,"end_s":680.64,"text":"either, is there? Which is already kind","speaker":null,"is_sponsor":0},{"start_s":679.04,"end_s":686.72,"text":"of an issue because it only ships with Wi-Fi 6E. So, you'd have to get like a","speaker":null,"is_sponsor":0},{"start_s":683.36,"end_s":690.04,"text":"Thunderbolt to PCIe to Wi-Fi 7 adapter","speaker":null,"is_sponsor":0},{"start_s":686.72,"end_s":692.279,"text":"card solution if you wanted faster","speaker":null,"is_sponsor":0},{"start_s":690.04,"end_s":697.0,"text":"Wi-Fi. Oh, Apple. Subscribe to ShortCircuit.","speaker":null,"is_sponsor":0}],"full_text":"You guys just saw our income breakdown, so you know that unless you buy an awful lot of Framework hoodies, I am not making back what I spent on this brand new M3 Ultra Max Studio. So, the question is, how can we possibly justify buying it? And the answer may surprise you. You see, depending on what you're using it for, the M3 Ultra Studio is actually, dare I say it, a good deal. Not for everyone, of course. If all you're doing is some light photo editing, you're going to be better off doing that on your phone. But if you work with AI, for instance, or if you have heavy demands when it comes to content creation, then even with our configuration, which has the topsp spec GPU and cost us about $10,000, but this little puppy could end up making you a lot more than you spent. Let's take a look after, of course, we uh look at the accessories. There's a power cable and paper. And that's it. So, after that very short side quest, let's take a closer look at the machine itself. Starting at the front, we've got an SDXC reader as well as two Thunderbolt ports. And these are one of the big differentiators for the new Mac Studio. Whether we're talking about the topsp spec M3 Ultra version that we've got here or the lesser M4 Max version that has faster cores, but not nearly as many of them. And the reason that these are such a big deal is because now they are Thunderbolt 5, which even under normal circumstances offers up to double the performance of last gen Thunderbolt 4 because it transitions from PCIe Gen 3 to PCIe Gen 4. That means more bandwidth for things like connected SSDs or card readers or external GPUs, at least on non-MAC machines. But it also has another trick up its sleeve. Instead of operating at 80 Gbit 80 Gbit transmission and receive, it can actually stack it and do up to 120 GB transmission and 40 GB receive in boost mode. This allows the new Mac Studio with its Thunderbolt 5 ports to drive extremely highresolution, high refresh rate displays, and a lot of them. Apple boasts it'll do up to eight connected displays in the ultra configuration. Now, let's take a look at the back where we find, that's right, another four of those Thunderbolt 5 ports, allowing you to connect pretty much whatever you could possibly want. remembering, of course, that Thunderbolt can daisy chain, so you can have up to six devices off of each of those six ports. We also get a 10 gig Ethernet connection, power input, and Mickey Mouse style, two USBA ports that operate at up to 5 GB per second, an HDMI port, as well as a headphone jack, and of course, conveniently located rear power button. We also get a nice beefy looking exhaust to go along with our nice beefy intake on the bottom. I haven't heard it in person, though I will very shortly. The Labs tells me is pretty darn quiet. A Kensington lock for the bottom. And that's about it. It's just a big block of aluminum. A big block of aluminum full of some very impressive specs. Let's talk speeds and feeds while I get this thing set up. The M3 Ultra SoC in here is equipped with up to 32 CPU cores. That is eight more than Apple's last gen flagship. But perhaps more importantly than that, all eight of the new cores are performance rather than efficiency cores, meaning this thing should be a multi-threaded beast. In addition to that, we get a new 32 core neural engine that is apparently much faster than last gen. And on top of that, we get up to an 80 core GPU with up to 800 gigabytes per second of combined memory bandwidth that is shared across all of those processors. This was a really key Apple innovation in the last few years, their unified memory. Of course, fast memory is great, but when it comes to AI, more memory is in some cases more important, and that is where this thing is really appealing to AI developers. It can be equipped with up to 512 GB of LPDDR5 memory. That means that compared to last gen, which topped out at 192, we can run much larger AI models. And that ignores the fact that the hardware is faster, allowing us to run them in a more performant manner. Of course, it's not all for AI. Apple loaded the M3 Ultra up with a new media engine that they claim supports up to 24 streams of ProRes playback in addition to supporting hardware encode and decode of many of the most popular formats, including AV1 decoding. Note that I said AV1 decoding, though. One major omission is that Apple's SOC's still do not support AV1 encoding. And this is true of both the M3 Ultra configuration as well as the M4 Max configuration. Either way, these machines are going to be Photoshop and Premier Powerhouses. Because even limited to CPUAV1 encoding, the M3 Ultra beat a Threadripper 7980X, a chip that consumes more power than this entire box and has twice as many cores. Pretty impressive. It's just worth noting that it doesn't support hardware encode, which is a shame when you look at the incredible performance of Apple's hardware encoder for H.264. Now, another thing that's worth noting is that Apple doesn't have a clear flagship with the Mac Studio anymore. I mean, they obviously have one that costs more, the M3 Ultra version, but the M4 Max version may end up outperforming it in some cases. Now, we don't have an M4 Max yet, but we do have an M4 Pro machine. And we found that in cases where singlethreaded CPU performance is more important than having a big beefy GPU or lots of CPU cores, it could end up being the more performant option. Of course, you'll also be limited to a maximum of 64 GB of memory. Now, before we go any further into performance, let's thank our sponsor, Dbrand. Uh, Dbrand told me they want you to go outside and touch grass. Okay, that shouldn't be that hard. I see some over there. I think there's some grass in there. There you go. Oh, sorry, boss man. I misread. They actually want you to touch grass on your Mac Studio. Oh, this. They actually told me they were working on this a while back. I didn't know they were ready to ship yet. That is really impressive. Apparently, every square centimeter contains a,000 grass strands. Man, you can really feel the density of it. It's like having a putting green on the top of your Mac Studio. This is not an April Fool's joke. This is a real product. Well, like the Touch Grass skin is a real product. They don't make skins for the Mac Studio, so get wrecked. I guess Dbrand has fantastic guides on their site for how to apply their skins, but I don't give a if those guys want me to follow their guide, they're going to have to pay me more and be a little nicer. There you have it, folks. This is as real of grass that you can stick to your phone. It's ultra limited edition. Once it's all sold out, it's sold out. You can get it now at shortlininus.com. All right, let's talk a little bit more about the Mac Studio, shall we? I want to fire up some kind of a load and see how loud this thing gets for myself. Let's go with Blender. We're going to do the Barberhop render. I'm going to see if I can hear this thing. I mean, surely it's got to ramp up. This is CPU mode, too. This is with all the CPU cores going full tilt. I can barely hear it. And I'm pretty much the biggest diva there is when it comes to computer noise. Like to the point where I put my computer in a separate room, so I don't have to hear it. If I could game on a Mac Studio, I probably wouldn't even bother. To be clear, I can game on a Mac Studio. Uh the performance is fine whether we're looking at games that rely on Rosetta or whether we're looking at games that run natively on Apple Silicon. But what it isn't is a tremendous value and the compatibility is still a major struggle for the macOS platform compared to Windows or even Linux these days. With that said, I still feel that GPU benchmarks are relevant because they give us some idea of what Apple has improved from generation to generation. And at least when it comes to gaming, it looks like not much. Most of this performance difference compared to the M2 Ultra could be explained by going from 76 to 80 GPU cores in the maximum configuration. But what's curious is that if I wasn't running this in CPU mode, if I was running Blender in GPU mode, it is much faster in GPU Blender rendering. And the same holds true in AI. Actually, in AI, performance is up across the board. So depending on the application, it'll use either the neural processor, the GPU, or the CPU, typically in that order in order to get the best performance out of the M3 Ultra SOC. And we found that across the board, the M3 Ultra is much faster than the M2 Ultra in AI. And that's in addition to being able to run much larger models. So, this was the main justification that Nick, and I'm naming him, Nick from the lab gave me for us buying this machine because we actually use our 192 GB M2 Ultra Studio daily to run local LLMs that help with programming for our coders. He says, and he promises me that this machine will justify itself because we'll be able to run better LLMs. Is that right, Nick? Yep. Okay. Oh, wow. I'm just looking a little closer at those gaming results. This thinging sucks for gaming. This is like on par with a 4060 Ti. The way it kind of keeps up in Blender, though, is so bizarre to me. Like, yeah, it's slower than a 7980X with an RTX 5090, but like not by that much, you know? Like a 4060 Ti would get crushed by that configuration, and this doesn't. What a strange strange machine. And boy, will you ever pay for it. The M4 Max configuration starts at $2,000. And the M3 Ultra starts at $4,000, going up as high as about $14,000 if you max out both the RAM like we did, and the storage, which we did not. A terabyte is fine when you're going to be accessing most of your storage over the network. Oh, with that said, could we maybe try to do a storage upgrade for an LT DIY? I don't know if there's a way to do that on this one yet. We'll have to figure that out because of course like other Apple silicon machines, Apple does not officially sanction any kind of storage or memory upgrade after the fact. You got to buy it and that's what you get. And actually, for that matter, I don't think there's any official way to upgrade like the the Wi-Fi chip in it either, is there? Which is already kind of an issue because it only ships with Wi-Fi 6E. So, you'd have to get like a Thunderbolt to PCIe to Wi-Fi 7 adapter card solution if you wanted faster Wi-Fi. Oh, Apple. Subscribe to ShortCircuit."}