naasking

4 ms·

31.
by naasking 1mo ago
Linking to X is less useful than it used to be.
32.
by naasking 1mo ago
Because nutritional value is not the only thing of value. We invented this clever way of uniformly valuing things called "money" that abstracts over all of those differences though.
33.
by naasking 1mo ago
> But that's tripled of water use across all sectors due to datacenter buildout. So no "comparison" to other uses is needed. You're saying almond farm water use tripled due to data center buildout? What other meaning
34.
by naasking 1mo ago
> Evidence actually supports that capabilities are leveling off What evidence?
35.
by naasking 1mo ago
My point is that generating is not illegal but sending is. Your whole point hinges on legality, so the distinction between legal and illegal seems pretty key.
36.
by naasking 1mo ago
Generating blackmail is not illegal, using it to blackmail someone is. Generating libel is not illegal, publishing it publicly is not illegal either although you can be sued over it. Generating worms and computer viruses is not illegal last
37.
by naasking 1mo ago
There is no question model quality is degraded by this though.
38.
by naasking 1mo ago
> Respectfully, tasks that allow for explicit straightforward true/false or done/not-done tests are not the "hardest real task[s]." Counterpoint: P vs. NP.
39.
by naasking 1mo ago
It's from work like this: https://arxiv.org/abs/2309.11495 A RL pipeline can reinforce verification behaviour even better than simple prompting.
40.
by naasking 1mo ago
> Yes. Actual real people believe that LLMs are actual, thinking, intelligences, perhaps even with consciousness. Yes, because it's possible. They don't have human consciousness/thought/intelligence, but it's ent
41.
by naasking 1mo ago
Yes, government regulations are almost always behind commercial entities making seemingly irrational choices.
42.
by naasking 1mo ago
I don't think "don't overthink" is a useful guide, it's too ambiguous which leads to the model overthinking how to interpret the advice to not overthink. I think guidelines have to be specific, such as "provide
43.
by naasking 1mo ago
To what extent can the overthinking be reduced by a better system prompt? For instance, telling it not to overthink, embellish or snazz anything up, and just directly deliver the most what was requested without in the plainest form to build
44.
by naasking 1mo ago
You'd need hundreds, if not thousands, of parts to justify a mold for casting. And even then, by the time you got the mold made you could have already printed hundreds of parts on a print farm, so you're losing time too.
45.
by naasking 1mo ago
Can anyone confirm whether this new Qwen release is any more concise when thinking? Overthinking was the biggest (only?) downside of the Qwen models.
46.
by naasking 1mo ago
You know free LLMs can help you understand that command line or design your own...
47.
by naasking 2mo ago
You need to do some reading: "Is raising mental health awareness causing more harm than good?", https://www.clinicaltrialsarena.com/analyst-comment/is-raisi... "Are We Overdoing Mental Health Awareness?&
48.
by naasking 2mo ago
No, in the past you just had the possibility of childhood illness and death, the death of siblings, war, famine, drought, none of which are real problems for university students these days.
49.
by naasking 2mo ago
What's less certain about modern environments compared to past ones?
50.
by naasking 2mo ago
The coddling is a big problem these days, and this focus on mental health is itself unhealthy. The assessment that "they would otherwise do well" is speculative, at best.
51.
by naasking 2mo ago
People say Qwen overthinks because they analyzed the thinking traces, and Qwen finds the answer relatively quickly but then second guesses itself multiple times for another 20,000+ tokens. Regardless of what other models do, that's cle
52.
by naasking 2mo ago
The continued vagueness suggests this isn't going anywhere, so I'll just conclude by saying that you've given nobody any real reason to believe this.
53.
by naasking 2mo ago
No such thing has been shown.
54.
by naasking 2mo ago
Typical LLM pretraining is very inefficient with data. NanoGPT slowrun shows that data can be used much more efficiently.
55.
by naasking 2mo ago
I don't understand all this dismissal of Einstein. What you just described is that he published a unified theory of Brownian motion, developed a new way of conceptualizing scientific theories via principles with which he derived specia
56.
by naasking 2mo ago
I use LLMs regularly for real work. I understand the current limitations, but you're talking about existing limitations that will never be overcome, because they cannot be overcome. I'm just asking for a couple of specific examp
57.
by naasking 2mo ago
Yes, it's worse because processors clocks are much faster than memory clocks.
58.
by naasking 2mo ago
Yes. That's why law is a different discipline from linguistics requiring its own rules of analysis, pedagogy, etc.
59.
by naasking 2mo ago
That's just vibes then. How is that supposed to be convincing?
60.
by naasking 2mo ago
> but there are many things which they are not good at which it is not cost effective or meaningful to improve Can you name a few such things so I can keep an eye on them in the coming years?

More ›