leonidasv

6 ms·

31.
by leonidasv 6mo ago
My 2 cents: Claude is more expensive, but it has something that Codex/GPT lacks that's not easy to quantify. Opus is probably a bigger model (my guess) and trained on code and technical writing (books?) of better qualify compared
32.
by leonidasv 6mo ago
And call MCPs as well
33.
by leonidasv 6mo ago
This is the same as saying "I still prefer hammer over screwdriver".
34.
by leonidasv 6mo ago
The comment system bundled with Wordpress is not enough for "serious blogging", you'll need plugins to manage spam, allow users to use social accounts, etc. You'll also need plugins for caching. Oh, and you'll need
35.
by leonidasv 6mo ago
It works on Safari.
36.
by leonidasv 6mo ago
Nice work and write-up! A side note: you embedded .mov videos inside <img> tags. This is not compatible with all browsers (notably Chrome and Firefox), which won't load the videos.
37.
by leonidasv 6mo ago
I've always wondered if the "typical" AI writing style is just an unavoidable RL artifact or a deliberate fingerprint to prevent model collapse as low-effort AI-generated text floods the training data pool (the web).
38.
by leonidasv 6mo ago
Yes and no, some techniques are still expensive for bots that aim to extract millions of addresses per day (like running JS and CSS, rendering SVG, etc).
39.
by leonidasv 7mo ago
Using Windows these days feels like "you're the product" in a way that's really annoying. Like, I know that "I am the product" when I open Instagram or Facebook, but the overall experience makes me not think ab
40.
by leonidasv 7mo ago
Congrats to the team. Unfortunately many comments here are missing the big picture by attacking the previous architectural decisions with no context about why they were taken. It's always easy to say so in retrospect. Also, I have to c
41.
by leonidasv 7mo ago
Those are compatible with the 1.x syntax while the gnata is compatible with the 2.x. Also, the repos haven't seen new commits in a long time.
42.
by leonidasv 7mo ago
> There's existing golang versions of jsonata, so this could have been achieved with those libraries too in theory The only one I found (jsonata-go) is a port of JSONata 1.x, while the gnata library they've published is compati
43.
by leonidasv 7mo ago
The link was changed, the old one did not mention it (apparently): https://news.ycombinator.com/item?id=47498220
44.
by leonidasv 7mo ago
Also, they were caught tampering saved webpages as well, so the website cannot be trusted to fulfill it's main purpose anymore: https://arstechnica.com/tech-policy/2026/02/wikipedia-bans-a...
45.
WebGL Considered Harmful (2011)(microsoft.com)
2 points by leonidasv 7mo ago|0 comments
46.
by leonidasv 7mo ago
Past discussion: https://news.ycombinator.com/item?id=12330899
47.
by leonidasv 8mo ago
Previous discussions: https://news.ycombinator.com/item?id=41895951 https://news.ycombinator.com/item?id=35716963
48.
The Part of PostgreSQL We Hate the Most (2023)(cs.cmu.edu)
3 points by leonidasv 8mo ago|1 comments
49.
by leonidasv 8mo ago
Stuxnet is also another mindblowing case. Wired write-up on it is a recommended reading: https://web.archive.org/web/20141028182107/http://www.wired....
50.
by leonidasv 8mo ago
If they're under an organised defamation campaign, they're not helping themselves by DDoSing someone else's blog and editing archived pages.
51.
by leonidasv 9mo ago
I use Claude Code daily to work on a large Python codebase and I'm yet to see the it hallucinating a variable or method (I always ask it to write and run unit tests, so that may be helping). Anyway, I don't think that's a pro
52.
Partly AI-generated folk-pop hit barred from Sweden's official charts(theguardian.com)
2 points by leonidasv 9mo ago|0 comments
53.
by leonidasv 9mo ago
There's also https://shademap.app/ for that, also useful (with 3D buildings!). Used it before buying my condo and it was spot-on.
54.
by leonidasv 9mo ago
Any way to run that on a Mac (besides running it in the browser)? $ ./snake.com ./snake.com: line 20: /tmp/a: cannot execute binary file
55.
by leonidasv 9mo ago
Unless we figure out how to make 1 billion+ tokens multimodal context windows (in a commercially viable way) and connect them to Google Docs/Slack/Notion/Zoom meetings/etc, I don't think it will simplify that much.
56.
by leonidasv 9mo ago
What profit? All labs are taking massive losses and there's no clear path to profit for most of them yet.
57.
by leonidasv 9mo ago
You don't need a LLM for that, a simple Markov Chain can solve that with a much smaller footprint.
58.
by leonidasv 9mo ago
> While some are still discussing why computers will never be able to pass the Turing test Are people still debating that? I thought it was settled by the time GPT-4 came out.
59.
by leonidasv 9mo ago
There's a trick shared here days ago to add a kind of shebang to Go that may interest you: https://lorentz.app/blog-item.html?id=go-shebang Discussion: https://news.ycombinator.com/item?id=46431028
60.
by leonidasv 9mo ago
I used to watch archived episodes of Computer Chronicles on YouTube almost every night before going to bed back in 2016~2018. It was my bedtime entertainment, watching those recordings from another era of computing and observing the hosts&#

More ›