crthpl

6 ms·

1.
by crthpl 13d ago
They were using it in May
2.
by crthpl 1mo ago
what specifically has changed about Claude Code?
3.
by crthpl 1mo ago
it does not use haiku. you can just ask Claude (possibly not fable BC of cyber) to reverse engineer the obfuscated JS.
4.
by crthpl 1mo ago
For the US models, you can look at the rewritten CoTs or just ask them what they think is happening.
5.
by crthpl 4mo ago
where the heck are you seeing OpenAI racing to IPO before anthropic?
6.
by crthpl 4mo ago
on the API, their margins are very high
7.
by crthpl 4mo ago
no, it is discounted expected future cash flows
8.
by crthpl 4mo ago
Yes it can! That's the whole point of RL! it generates slightly out of distribution rollouts, and rewards good rollouts to change the distribution of the output
9.
by crthpl 4mo ago
where the heck did you get those parameter numbers from?
10.
by crthpl 5mo ago
are you sure it's 60 fps and not uncapped?
11.
by crthpl 6mo ago
have you tried pangram? it's basically the only good AI detector, and they have nearly 0 false positives
12.
by crthpl 7mo ago
Anthropic is very big (the biggest AI co?) in B2B, where you don't have ads. Also, if they end up creating a datacenter full of geniuses, ads won't make sense either.
13.
by crthpl 8mo ago
1. you can definitely tell apart an S-curve and an exponential if you look at the derivative(s). AI progress does not seem to be close to the middle of the S-curve. 2. e.g.: a slowdown hasn't shown up in moore's law yet.
14.
by crthpl 8mo ago
It is the case that Anthropic employees have no usage limits. Some people do experiments where they spawn up hundreds of Claude instances just to see if any of them succeed.
15.
by crthpl 9mo ago
the chain of thought is what it is thinking
16.
by crthpl 1y ago
The reason they get a perfect score on AIME is because every question on AIME had lots of thought put into it, and it was made sure that everything was a possible. SWE-bench, and many other AI benchmarks, have lots of eval noise, where ther
17.
by crthpl 1y ago
GPT-4o is so far behind the frontier; you shouldn't use it as an indicator of what LLMs are capable of.
18.
by crthpl 1y ago
Another reason may be that WebGPU didn't allow for as much optimization and control as Vulkan, and the performance isn't as good as Vulkan. WebGPU also doesn't have all the extensions that Vulkan has.
19.
by crthpl 1y ago
If humans were too complicated to be mathematically modelled, then we wouldn't exist.
20.
by crthpl 2y ago
Why would charging for extended support hurt Microsoft? They "win" if you pay for it, or if you upgrade to windows 11. If you do neither, you'll probably get hacked down the line if you don't airgap, and Microsoft doesn&
21.
by crthpl 2y ago
I think there is a deterministic algorithm (based on GPS position?) for determining who climbs and who descends.
22.
by crthpl 2y ago
I'm guessing MacOS, Linux, Android, Windows, iOS, and Web?
23.
by crthpl 3y ago
The idiom "Don't judge a book by its cover" applies to many situations, but books are not one of them.
24.
by crthpl 3y ago
dawn is the WebGPU backend in chromium, while wgpu is the WebGPU backend for firefox written in Rust. wgpu is seeing a lot of use in non-browser uses; there are some examples on their website. https://wgpu.rs/
25.
by crthpl 3y ago
The 7 million is probably how many different hues we can see. We can see many more different brightness levels.
26.
by crthpl 3y ago
Not if only one species came to earth.
27.
Framework Laptop 16 Deep Dive – Enclosure(frame.work)
307 points by crthpl 3y ago|232 comments
28.
by crthpl 4y ago
Or perhaps the other way around: we are social because we have language.
29.
by crthpl 4y ago
Their Performance section shows that the 12GB is 50%-150% faster than the 3080 Ti (so better than 3090 Ti!), and it also has more cuda cores and memory than the 3070.
30.
by crthpl 4y ago
Eye candy is one of the main reasons people build Lego bulids.

More ›