ramon156

4 ms·

31.
by ramon156 16d ago
good thing US is completely clean
32.
by ramon156 16d ago
what a fabled way of thinking
33.
Anthropic Researcher Quits over 'Out-of-Control' AI Fears(wsj.com)
3 points by ramon156 16d ago|0 comments
34.
by ramon156 17d ago
Reminds me of the first season of Succession
35.
by ramon156 18d ago
They have their own infra but its not the size of CF. Chicken and Egg
36.
by ramon156 18d ago
big corps = big money. they dont make money from your $5/mth static site, its practically free when you're at CF scale
37.
by ramon156 19d ago
In the netherlands we have 9292 which imo is the most mature implementation.
38.
by ramon156 19d ago
Disclaimer: I hand curate them in the end I keep most of my sessions in Zed (you can import them there anyway). After some big feature I let a frontier agent go over these sessions and suggest improvements. Typically I use gemini for this b
39.
by ramon156 20d ago
I didnt even need to open the link, a ghibli-generated profile pic already says so much about someone's priorities.
40.
by ramon156 21d ago
orr, ~/.gitignore
41.
by ramon156 21d ago
great idea, sadly full of vibe-code. I assume it was prompted by copying the style of some existing EU-style Design system. Idea is promising though, hope it matures into something solid.
42.
by ramon156 21d ago
Both OAI and Anthropic seem to have released a model that is slightly better but cost ~2x the previous iteration. Interesting play
43.
by ramon156 21d ago
Do parents feel safer with these checks in place? Kids aren't dumb, and its pretty clear these verifications can be bypassed. is it just to appear good?
44.
by ramon156 22d ago
I'm somewhat pessimistic about AI tools, but I quite like Shopify's approach. Sidekick is pretty useful for merchants who do not want to navigate the UI, and their IMG generation is pretty good for demo stores. I'd like to se
45.
by ramon156 22d ago
how come we shifted away from tree-sitter? i still feel like if you're going to let a neural network program, let it generaye an AST. it doesnt matter if it gets the name wrong, as long as the code makes sense.
46.
by ramon156 22d ago
i love these write-ups. keeps you up to date while not needing to dig through PRs
47.
by ramon156 22d ago
blegh, both the code comments and readme reek of verbose claude puke. > The registered Lean build passed without errors or warnings. thanks. i like it when my proofs do not error. Edit: just saw this was published under openai, yikes.
48.
by ramon156 23d ago
so local, but with cors, api keys and remote calls? who is this for? people who have a powerful pc but want nothing on their own device?
49.
by ramon156 23d ago
Love what Audacity has become, especially the new UI that does not look like it was forced through some scrum pipeline
50.
by ramon156 23d ago
something that i realisr with LLM-generated content is that I just get more confused every sentence. HOW does it work? WHAT does it do? When I see technology I want to figure out what it CAN'T do. For example, Rust can do web, but I wo
51.
by ramon156 24d ago
The only company with a proper TPU set-up is bound to have the fast models, now add a market cap like Google to the mix.
52.
by ramon156 24d ago
The way I read it, I thought they would just check if the file hash exists on their end. Is that not the case?
53.
by ramon156 25d ago
what did i read? genuinely? its only a few words, and even that had to be AI written. The topic in the title barely was mentioned.
54.
by ramon156 25d ago
I yearn for a model that can churn through claude text and write sensible text. so far gemini is pretty good at that, even in the low variant
55.
by ramon156 25d ago
when tools call themselves "blazingly-fast", at least a benchmark can prove that. calling your own product comprehensive is odd.
56.
by ramon156 25d ago
this website's intro to a transformer is longer than just reading the paper itself
57.
by ramon156 25d ago
Every feature of Atlassian feels like a KPI needed to be hit as opposed to making quality software
58.
by ramon156 25d ago
the usage and quality out of gpt 5.6 is on par, if not better in token usage. I would love to do a write-up on which tools are useful for what use-case, as AA keeps improving. I use gemini for rewriting code docs, because the frontier model
59.
by ramon156 25d ago
I'd love to put in some love and effort into a Spotify replacemeny that supports jams. I'd share the app with friends. I needed an excuse to make a cross-platform app anyway.
60.
by ramon156 25d ago
really cool write-up! not so sure about the "2600 skills". is mors really better? are LLMs " good enough" at not overfitting their context?

More ›