Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
zaptrem
searching PlanetScale…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
15 ms
·
151.
▲
by
zaptrem
2y ago
I used a Surface Pro for 6 years and and haven’t missed the touch screen once since switching back to MBP 3 years ago. I would have missed the handwriting input but that’s what a low end iPad is for.
152.
▲
by
zaptrem
2y ago
Sorry what is the thing I want in this case? No stuttering or no crashing?
153.
▲
by
zaptrem
2y ago
This happens whenever I load up one of our PyTorch models on my M1 MBP 16gb too. I also hate the part where if the model (or any other set of programs) uses too much RAM the whole system will sometimes straight up hang and then crash due to
154.
▲
by
zaptrem
2y ago
To be clear, my use of the word “photorealistic” instead of “accurate” was very intentional.
155.
▲
by
zaptrem
2y ago
Even with antibiotics strep really sucks for kids (painful!) and parents (kid needs to stay home from school). Seems like there would be obvious demand. The article says the FDA banned testing for like 40 years due to a tiny issue during a
156.
▲
by
zaptrem
2y ago
This isn’t true in any way we could reasonably reach. The world has an insane amount of potential food output. A huge portion of the food the US produces goes directly in the trash instead of being used.
157.
▲
by
zaptrem
2y ago
While this is understandable, the unfortunate issue was that Apple could be coerced into adding images certain authoritarian governments didn’t like to the list. Though imo it’s all moot if iCloud Photos aren’t end to end encrypted anyway.
158.
▲
by
zaptrem
2y ago
Wouldn't that require exfiltrating the original photo? I remember them swearing that wasn't part of the deal.
159.
▲
by
zaptrem
2y ago
https://anishathalye.com/inverting-photodna/ You could almost certainly produce nearly photo realistic PhotoDNA inversions with a finetuned diffusion model now. Is it possible to create a perceptual hashing algorithm w
160.
▲
by
zaptrem
2y ago
AI gets backpropagation whereas evolution is more like particle swarm optimization (pick a bunch of random values, then pick more random values near the ones that do best). Backpropagation is way better/faster since in expectation the
161.
▲
by
zaptrem
2y ago
Exactly what the labels want, since if that type of thing keeps going their way they will soon own not just every song like they currently do, but every future song forever.
162.
▲
by
zaptrem
2y ago
Copyright means you own the right to reproduce a given work. It doesn't mean you own the ideas behind that work. If that were true, then all of modern music would instantly be a copyright violation.
163.
▲
by
zaptrem
2y ago
Yeah, I wonder whether this still saves compute if you include the compute used to train DINOV2/whatever representation model you'd like to use?
164.
▲
by
zaptrem
2y ago
> No matter where you live, you’ll always be able to exercise your right to opt out of this data sharing by updating your preference settings in your account under “Data and Privacy.” There is currently no opt out button or switch on thi
165.
▲
by
zaptrem
2y ago
Running preprocessing jobs on a $0.5 SFCompute H100 node RN (though price usually bounces up to what you mentioned).
166.
▲
by
zaptrem
2y ago
As someone who has done a bunch of large scale ML on hyperscaler hardware I will say the uptime is nowhere near 99.9999%. Given a cluster of only a few hundred GPUs one or multiple failures is a near certainty to the point where we spend
167.
▲
by
zaptrem
2y ago
Don’t forget Apple’s GamePortingToolkit based on Crossover/Wine and the open source client for it Whisky. I think it supports most games Linux Proton does now.
168.
▲
by
zaptrem
2y ago
World Labs’ (Lee’s company) stated goal is to “[build] Large World Models to perceive, generate, and interact with the 3D world.” Imo video models are the closest thing we have to “spatial intelligence.” They generate in three dimensions (2
169.
▲
by
zaptrem
2y ago
When using two way Bluetooth the bitrate drops sharply. No amount of cleaning can solve this.
170.
▲
by
zaptrem
2y ago
Driving has a lot more visual/motion-based stimuli than a Mission Control terminal (assuming that's the setting they were testing with).
171.
▲
by
zaptrem
2y ago
> Wifi7 has 3-D radar features for gestures, heartbeat, keystrokes and human activity recognition, which requires the NPU inside Intel SoC. The M.2 card is only a subset. Source? Google turned up nothing.
172.
▲
by
zaptrem
2y ago
Google still has an unbelievable training infrastructure advantage. The second they can figure out how to convert that directly to model performance without worrying about data (as the o1 blog post seemed to imply OAI had) they’ll be kings.
173.
▲
by
zaptrem
2y ago
Classic "code and weights released at X." But when you go to the repo at X there's nothing there and possibly never will be.
174.
▲
by
zaptrem
2y ago
I don't have to use LinkedIn and they can't throw me in jail or worse. Most people have no choice but to use their government (yeah I guess I could move but most cannot and it assumes there's another country that would take y
175.
▲
by
zaptrem
2y ago
By default you do not have a LinkedIn account.
176.
▲
by
zaptrem
2y ago
Should we kill the Internet Archive too?
177.
▲
by
zaptrem
2y ago
Many games that work in Wine work just fine in GamePortingToolKit (which is just an Apple-written patch on top of CrossOver which is just more patches on top of Wine)
178.
▲
by
zaptrem
2y ago
I see the open weights/source model community as the ones willing to give the baby away.
179.
▲
by
zaptrem
2y ago
> Betterment of humanity? Blink twice if a LinkedIn PR person is holding you at gunpoint I was referring to Generative AI in general, this use-case is quite boring. > a corporate behemoth vacuuming people's personal information s
180.
▲
by
zaptrem
2y ago
I understand and completely agree republishing content (even altered) isn't cool, and I also agree government use of technology for mass surveillance is incompatible with our idea of democratic/open societies. However, in the
More ›