m11a

7 ms·

1.
by m11a 7d ago
Yes it's definitely not SOTA, and is worse than all the chatbots I use. But it's correct frequently enough that the convenience + speed does it for me.
2.
by m11a 8d ago
Although there are companies trying to work around that too, from PhysicsX to some of the world model co’s.
3.
by m11a 8d ago
tbh I also use google AI tools frequently. It’s convenient (search bar is top of my window), and extremely fast at generation. For quick Qs, opening a chat bot and changing the default reasoning from my coding settings, switching from Work
4.
by m11a 13d ago
Many VCs also dislike these examples, I believe. I'm doubtful this is what they're being pitched. As for public releases: I wonder if it's because these examples are easy to relate to. Many websites are just a long tail of in
5.
by m11a 13d ago
I recall them saying they use models to write CUDA kernels and whatnot. Makes sense, and unsurprising that models are good at writing code. But I think calling this “automating AI research” is misleading. I’m not sure there’s evidence yet t
6.
by m11a 13d ago
I did, at least until Fable 5.1. Grok’s models are excellent, amazing price-performance and speed too.
7.
by m11a 2mo ago
In essence. But if you have a chain PR1 > PR2 > PR3, and PR1 gets merged, all the others (ime) seem to not cleanly rebase on main. They end up with conflicts that require manual fixing. I've not really figured out why, tbh. It&#x
8.
by m11a 2mo ago
GitHub's PR workflow doesn't nicely support being able to review individual commits, realise which comments are associated with which commits, etc. Or shipping individual commits to main, while working on some others (unless you a
9.
by m11a 2mo ago
I think it's telling how long it took GitHub to release a v1 of this feature. Folks have wanted this for a long time. Graphite came along and did it years ago (and I'm sure they pondered whether GitHub would do this). And the v1 i
10.
by m11a 2mo ago
I presume such US legislation isn't going to try claim worldwide jurisdiction to block all persons worldwide from using Chinese models. In which case, the French guy wouldn't be violating American law. As for the American company,
11.
by m11a 2mo ago
Because it's not losing money on each token? Aside from most global people using American inference providers to run the models, I suspect the cloud inference products of the Chinese labs are profitable, at least on the inference costs
12.
by m11a 2mo ago
Probably just the US. But the US could do what EU has done with e.g. GDPR, Digital Services Act, USB-C regs, where they force any company trading in their region to follow those regulations for domestic customers. And basically any AI compa
13.
by m11a 2mo ago
Step 1: Chinese company publishes open weights on HF Step 2: European company distills or just adjusts the model slightly, and publishes its model on HF Step 3: American company uses model from step 2. Has to testify under oath where they g
14.
by m11a 2mo ago
How long did it take from the first DBMS to get to Postgres? The first OS to get to Linux? The first compiler to get to LLVM? For Postgres and Linux and LLVM to become mature enough to hold the revered reputation they have now? The jury&#x2
15.
by m11a 2mo ago
I had conversations with Ben a few years ago about his economics research (incidentally related to some research I was doing at the time). I wouldn't have guessed he was close to 70. He's a sharp guy. (I also think his economics r
16.
by m11a 2mo ago
Right. But given that the entire point revolves around QuePaxa, it's strange to see no discussion on it. If that weren't the point, the article would be "Why Cloudflare implemented and deployed Paxos". Which would also b
17.
by m11a 2mo ago
This article is a bit hard for me to grasp the main ideas of because, given Cloudflare's requirements (e.g. no strong leaders), it immediately seems like they should be comparing to leaderless protocols like Paxos-class algorithms. Com
18.
by m11a 2mo ago
It depends on the model provider. OpenAI's is very limited and precisely written. Plus, open-source models hosted on SaaS inference providers tend to come with a strong ZDR agreement too.
19.
by m11a 2mo ago
Most corporations likely have zero data retention agreements with LLM providers, at least for API usage. (Sure, you could be sceptical on whether the LLM provider is upholding that, but I personally do trust them. The trust betrayal if ZDR
20.
by m11a 3mo ago
Says a lot about just how much money they’re raking in from core business.
21.
by m11a 3mo ago
I mean, you’d resort to an obfuscated approach if you thought the ‘malicious’ users would remove your direct telemetry. The other users could be perfectly happy about it, but if you announced the change, obviously the malicious users would
22.
by m11a 3mo ago
I mean, after the US just signed an export control ordering Fable access blocked to non-US users (including European nationals), I doubt European and "US-aligned markets" are eager to ban Chinese models against their own interests
23.
by m11a 4mo ago
California already has very high taxes. I think marginal tax rates are higher in California than for UK tax residents, certainly for CGT, and roughly similar for income tax. I'd say the fact that California remains the epicenter of tec
24.
by m11a 4mo ago
If Railway's account was suspended due to an error, not a TOS violation, I doubt they'll pull such a card. If Google were sued, such a blatant lie would be found in discovery pretty quickly and I doubt a court would look upon it f
25.
by m11a 4mo ago
I’ve learned that in AI, until you can get your hands on something and try it yourself, assume all claims are false.
26.
by m11a 5mo ago
I'm not sure security through obscurity is a great practice? Not to mention, I presume the core bits of Cal.com's source code are already in place and aren't going to change significantly? Like, this feels like a business dec
27.
by m11a 6mo ago
Cloudflare's Code Mode is conceptually the same as Anthropic's Code Mode ( https://www.anthropic.com/engineering/code-execution-with-mc... ), or the various open source implementations that predate and postdate
28.
by m11a 6mo ago
If I recall correctly, the ‘official’ Python one is a fork of FastMCP v1 (which then removed the attribution, arguably in violation of the original software’s license)
29.
by m11a 6mo ago
K8s gives you orchestration of Docker containers. I don’t think it handles the container boundary any more than Docker does. I don’t think it should be assumed to give network isolation, unless you’re also using extensions and something lik
30.
by m11a 7mo ago
Don't think that's a valid comparison. Apple can do those things because they control the hardware device, which has physical distribution, and they lock down the ecosystem. There is no third party app store, and you can't ge

More ›