bionhoward

6 ms·

1.
by bionhoward 13d ago
I think the API runs on Azure
2.
by bionhoward 13d ago
I personally think this acquisition is genius anybody who could afford it would be wise to own HuggingFace because they are incredible
3.
by bionhoward 1mo ago
I was wrong here, it’s more about the RNG they use for the word choices [1]. Although I wouldn’t be surprised to see UTF-8 substitutions also. [1] https://www.anthropic.com/news/claude-text-watermark
4.
by bionhoward 1mo ago
I think it’s a bunch of hidden UTF-8 character substitutions so the text looks the same but the underlying bytes are different
5.
by bionhoward 3mo ago
It’s amazing how much Python punishes you for modularizing your code
6.
by bionhoward 4mo ago
I’m heavy into rust and never really use golang, but one big benefit of go over rust is compile times are significantly quicker, which could be more fun if you’re running CI checks 50 billion times
7.
by bionhoward 5mo ago
Yeah it’s a coding beast if you get it dialed in right
8.
by bionhoward 6mo ago
gotta admit i groaned a bit at this because it would make rust more complicated, but on my 2nd read i realized: - some things (compile time bounds checking tensor shapes) are hard / impossible to implement now; "pattern types&quot
9.
by bionhoward 6mo ago
Anthropic is not trustworthy for this because they force every Claude Code user to agree to a noncompete while also opting them in to model training. That means, by default, every Claude Code user is actively getting royally screwed
10.
by bionhoward 7mo ago
surprisingly accurate! Is Gary the AI equivalent of the “nothing ever happens” guy?
11.
by bionhoward 7mo ago
genius, great idea and follow through, please keep it up, this could improve the ML industry tremendously, maybe some einops inspired interface for this would be good?
12.
by bionhoward 7mo ago
building this openclaw thing that competes with openai using codex is against the openai terms of service, which say you can't use it to make stuff that competes with them. but they compete with everyone. by giving zero fucks (or just
13.
by bionhoward 7mo ago
signal should just add better API / bot stuff and then we could all use that. there's no way OpenAI would be trustworthy for this; slack certainly isn't
14.
by bionhoward 8mo ago
found a bug searching for "collagen": Failed to execute 'removeChild' on 'Node': The node to be removed is not a child of this node. Something broke. We're working on it, but in the meantime, try rel
15.
by bionhoward 8mo ago
Yes! The paywalled SQL documents are a big annoyance
16.
by bionhoward 8mo ago
Yup, I’ve been crowing about these customer noncompetes for years now and it’s clear Anthropic has one of the worst ones. The real kicker is, since Claude Code can do anything, you’re technically not allowed to use it for anything, and ever
17.
by bionhoward 8mo ago
Why act like it’s a mystery when the Claude Code repo clearly explains: > When you use Claude Code, we collect feedback, which includes usage data (such as code acceptance or rejections), associated conversation data, and user feedback s
18.
by bionhoward 9mo ago
Would SRAM make weight updates prohibitive vs DRAM?
19.
by bionhoward 9mo ago
Gemini is absolutely not Gryffindor since it auto-opts users into training AI on their codebases without informed consent
20.
by bionhoward 10mo ago
Anthropic also has legal terms that say no one is allowed to use the service for anything work related, but nobody seems to care
21.
by bionhoward 10mo ago
Cursor still wins over Claude Code because Cursor has privacy mode
22.
by bionhoward 10mo ago
I think it’s more about deciding how much to think about stuff and not a model router per se. 5 and 5.1 get progressively better calibrated reasoning token budgets. Also o3 and “reasoning with tools” for a massive consumer audience was a ma
23.
by bionhoward 10mo ago
model provider CLIs are a trap, less freedom of choice, less privacy, way more prohibitions buried in the fine print
24.
by bionhoward 11mo ago
Good reason to use Cursor, you can insta-switch to whatever model you want and even run diverse models from different providers at the same time. If one of em isn’t working then you can try something else instead of being stuck on one model
25.
by bionhoward 11mo ago
And here I thought it was a pedantic word for “data box”
26.
by bionhoward 11mo ago
Is using these terminal agents with customer noncompete and no privacy questionable when cursor has the same models and privacy mode?
27.
by bionhoward 11mo ago
The point is it’s sucking your data into some amorphous big brother dataset without explicitly asking you if you want that to happen first. Opt out AI features are generally rude, trashy, low-class, money grubbing data grabs
28.
by bionhoward 1y ago
Rust for everything, I could explain but it would get rude
29.
by bionhoward 1y ago
Rust feels easy for me, could it be we’re just used to what we use more? Anyway, it’s all pretty easy, what’s the use arguing which of multiple easy things is easiest?
30.
by bionhoward 1y ago
Wouldn’t a breakpoint or log message require the code to compile and run in order to work?

More ›