khalic

13 ms·

331.
by khalic 7mo ago
Some are stuck in 2010s, where people thought that JS was turning into a lingua franca. As usual, such delusions are costing us some pretty heavy price. People seem to now accept crappy, laggy UIs "because it makes business sense"
332.
by khalic 7mo ago
What's this armchair lawyer interpretation I'm hearing these last weeks, "LLM output doesn't seem protected by copyright"? It's extremely clear, from jurisprudence, that the level of human intervention in the p
333.
by khalic 7mo ago
Fair points, I didn’t know about the legal tango from the TDF, circumventing processes to impose yourself is not the tool of the righteous usually
334.
by khalic 7mo ago
Full circle indeed, nice historical capsule thanks
335.
by khalic 7mo ago
The use case is well defined here, let’s not jump the gun. Text search, like with code, is a relatively simple problem compared to intrinsic semantic content in a book for example. I think the moral here is that RAG is not a silver bullet,
336.
by khalic 7mo ago
This kind of circles back to ontological NLP, that was using knowledge representation as a primitive for language processing. There is _a ton_ of work in that direction.
337.
by khalic 7mo ago
This is anecdotal at best, but it does play into the tired old technical vs non-technical simplification. The fact that the two entities have now become direct competitors is a better explanation grounded in facts
338.
by khalic 7mo ago
They’re relaunching Libre Office online apparently, they don’t want competitors on their board I’m guessing
339.
by khalic 7mo ago
So… a prompt? I’m not on my laptop but I hooked it to cmp.nvim, gave it a short situational prompt, +- 10 lines, and started typing. Not anywhere near usable but with a little effort you can get something ok for repetitive tasks. Maybe some
340.
by khalic 7mo ago
So, basically, TDF doesn’t want Collabora (a company) people on their board. The technical vs non-technical framing seems contrived at best. The excuse by TDF seems… suspicious.
341.
by khalic 7mo ago
You can hide malicious intent, so the repeated negligence patterns you’re pointing out make a better signal. Smart. Thx for the perspective
342.
by khalic 7mo ago
If you’re looking into small models for tiny local tasks, you should try Qwen coder 0,5B. It’s more of an experiment, but it can output decent functions given the right context instructions.
343.
by khalic 7mo ago
Making a sentence out of a json
344.
by khalic 7mo ago
It’s not like creating a chip gives you unfettered access to it. You _can_ add 0-day flaws and backdoors, but these can be discovered, leaked, etc. Has there been any case of such a backdoor built in consumer chips like theses? I’m not talk
345.
by khalic 7mo ago
AFM models are very impressive, but they’re not made for conversation, so keep your expectations down in chat mode.
346.
by khalic 7mo ago
Where? Any war games in the last 10 years. It’s a known issue with aircraft carriers agiants anti ship missiles. What’s protecting them right now is what would happen to a country if they attacked one of those. Retribution is not a great de
347.
by khalic 7mo ago
100% interception … drone interception is NP complete dude, there’s nothing you can do against 1000 drones like that, and they’ll get cheaper, faster, smaller, bigger, more manoeuvrable. So 10Million bucks to down an aircraft carrier. With
348.
by khalic 7mo ago
Lol carriers were already being overwhelmed by regular missiles, this now means a multi billion dollar ship can and will be destroyed by cheap drones if it's anywhere near its optimal deployment zone.
349.
by khalic 7mo ago
This study assumes everybody is oblivious to contamination, and explicitly says they can't differentiate. Not useful and bordering on the tautological
350.
by khalic 7mo ago
Misleading title
351.
by khalic 7mo ago
Just… keep coding?
352.
by khalic 8mo ago
Really cool trivia, thanks for sharing :D
353.
by khalic 8mo ago
Totally realistic expectation
354.
by khalic 8mo ago
Keep wishing, in the meantime some people have to deal with the real world and plan accordingly
355.
by khalic 8mo ago
... you think It was good time? Not being able to publish anything without sifting through all the libs licences? Remembering legalese, jurisprudence, edge cases, on top of everything else? MIT became ubiquitous because it gives us peace of
356.
by khalic 8mo ago
Unenforceable means they can't actually enforce it since they can't discriminate high quality LLM code from hand typed
357.
by khalic 8mo ago
Sarcasm? Nobody will be contributing with a complexe signing process like that, and it doesn't guarantee anything in the end, it's like a high tech pinky swear
358.
by khalic 8mo ago
The LLM ban is unenforceable, they must know this. Is it to scare off the most obvious stuff and have a way to kick people off easily in case of incomplete evidence?
359.
by khalic 8mo ago
They’re all muscle inside, fascinating
360.
by khalic 8mo ago
I see the industry found its new buzzword: humanoid robots Slap a "head" on an industrial machines and watch investors go brrrrrrrrr

More ›