XCSme

32 ms·

571.
by XCSme 10mo ago
Good point. It also reminded me of when I was trying to optimize my app for some scenarios, then I realized it's better to optimize it for ALL scenarios, so it works fast and the servers can handle no matter what. To be more specific,
572.
by XCSme 10mo ago
I've recently added error tracking to my self-hosted analytics app (UXWizz), and the way I did it is simply add extra events to each user/session. Once you have the concept of a session or user, you can simply attach errors or log
573.
by XCSme 10mo ago
I hope they'll release a new AM4 CPU Something like 5900x on 2nm or 4nm
574.
by XCSme 10mo ago
Awesome! Is it actually working well? Not really, at least not at this stage. But it's cool to see a new UX idea.
575.
by XCSme 10mo ago
It's working now.
576.
by XCSme 10mo ago
Anyone else getting ERR_TOO_MANY_REDIRECTS trying to access the post?
577.
by XCSme 10mo ago
LLMs are weird, Gemini 3 flash beats Gemini 3 Pro on some benchmarks (MMMU-PRO)
578.
by XCSme 10mo ago
Just a few more credits and it will finally fix that bug without introducing new ones, exactly how I asked
579.
by XCSme 10mo ago
It's also laggy for me, with a 5900x + 3090...
580.
by XCSme 10mo ago
The demo is not loading for me: https://i.snipboard.io/gkNxDO.jpg Probably because of: https://i.snipboard.io/XDQUWG.jpg
581.
by XCSme 10mo ago
Contracting work doing n8n AI automations. My main personal project is a self-hosted Hotjar alternative: https://www.uxwizz.com
582.
by XCSme 10mo ago
I would assume a good benchmark has hidden tests, or something randomly generated that is harder to game
583.
by XCSme 10mo ago
But most benchmarks are not about that... Are there even any "hallucination" public benchmarks?
584.
by XCSme 11mo ago
Maybe they are just more consistent, which is a bit hard to notice immediately.
585.
by XCSme 11mo ago
Neurons are huge
586.
by XCSme 11mo ago
Oh, tensorflow. When AI actually meant creating your own networks...
587.
by XCSme 11mo ago
I think LLMs are quite good at summarization tbh, usually very accurate for articles this type of length.
588.
by XCSme 11mo ago
Why is downvoted? Is it wrong? It does continue to say it's just a short-term burst and bad in the long-term.
589.
by XCSme 11mo ago
Probably not worth the effort, for a couple days of downtime, we'll just move somewhere else. But I agree, I recognize the silence in that forum thread that was locked without a resolution: some boss said "let they complain or pay
590.
by XCSme 11mo ago
I am curious if there are any local-model grammar tools out there? There are many, good-enough tiny models that could technically run in the browser.
591.
LanguageTool browser extension is no longer free(languagetool.org)
2 points by XCSme 11mo ago|1 comments
592.
by XCSme 11mo ago
In my experience building AI Agents, having more specialized agents works better than one big agent. Simply because LLMs are dumb and make A LOT of mistakes. The longer the system prompt and more instructions you give it, the more likely it
593.
by XCSme 11mo ago
That's a cool work-around. What I don't like, is that whenever you contact Zoom, their representatives are taught to say one thing: buy more licenses. Not only that, but their API/pricing is specifically designed to cover edg
594.
by XCSme 11mo ago
Artifical limits, because they have 40 paid licenses that they can not use, because of a non-disclosed assignment limit that is NOT mentioned in the pricing page nor in the ToS. A lot of people are angry about this, and I think it's bo
595.
by XCSme 11mo ago
Today a client is having some issue with Zoom because of some artificial rate limits they impose. Their support is not responding, the account can't be used, courses can not be held and there's not much we can do. We already start
596.
by XCSme 11mo ago
I mean, I still prefer to have the ability to fix it myself, because I know I can probably do it in 1h max. I know this doesn't apply to most people, especially those outside of HN though.
597.
by XCSme 11mo ago
Plus, when you self-host, you can likely fix the issue yourself in a couple of hours max, instead of waiting indefinitely for a fix or support that might never come.
598.
by XCSme 11mo ago
If it's backtesting on data older than the model, then strategy can have lookahead bias, because the model might already know what big events will happen that can influence the stock markets.
599.
by XCSme 11mo ago
If you are a smaller business, you don't need those enterprise solutions, something like self-hosted PostHog/UXWizz/Matomo should be more than enough. No reason to send your data to other companies.
600.
by XCSme 11mo ago
Such a big company should be able to easily self-host their analytics. They don't even have to create their own platform, there are many out there that they can use.

More ›