naasking

26 ms·

331.
by naasking 9mo ago
Your argument just assumes there is no latent structure that can be exploited. That's a big assumption.
332.
by naasking 9mo ago
Indeed, and I think natural language and reasoning will have some kind of geometric properties as well. Attention is just a sledgehammer that lets us brute force our way around not understanding that structure well. I think the next step ch
333.
by naasking 9mo ago
I think any kind of innovation here will have to take advantage of some structure inherent to the problem, like eliminating attention in favour of geometric structures like Grassman flows [1]. [1] Attention Is Not What You Need, https:&#x2
334.
by naasking 9mo ago
I'm not sure anonymous users should be able to join. Arxiv's system of only allowing academic users seems fine for this, although exceptions could be made for industry researchers.
335.
by naasking 9mo ago
The LLM confusion is just the latest incarnation of the Confused Deputy problem. It's in the same class of vulnerabilities as CSRF.
336.
by naasking 9mo ago
Calling it peer review suggests gatekeeping. I suggest no gatekeepind just let any academic post a review, and maybe upvote/downvote and let crowdsourcing handle the rest.
337.
by naasking 9mo ago
> AI software engineering in a nutshell. Leaving the human artisan era of code behind. Function over form. Substance over style. Getting stuff done The invention of calculators and computers also left the human artisan era of slide rules
338.
by naasking 9mo ago
> This is not correct. There are clear instructions on how a 4-way stop should operate and its yielding to the right Yielding to the right only applies if you stop at roughly the same time, otherwise first to stop goes first. It's t
339.
by naasking 9mo ago
CERN and large construction projects like the FCC employ tens of thousands of physicists and engineers across decades. It's hard to convince someone of something when their livelihood depends on them not believing it.
340.
by naasking 9mo ago
CERN shouldn't get that funding if they're going to push for the FCC. Funding is often contingent. Even sticking only to particle physics, there's lots of other research that could use that funding better, like wakefield acce
341.
by naasking 9mo ago
Payoffs have many forms, the most important for pure research being "advancement of knowledge". We have nearly zero expectation of knowledge advancement from yet another radio frequency collider.
342.
by naasking 9mo ago
> It is hard to gauge this is in advance though. If you were sure what you were gonna find, it wouldn't be much of a discovery. Virtually all previous particle discoveries were predicted, and then we built devices to find them, eg.
343.
by naasking 9mo ago
1. That's not an argument unless the evidence for these payoffs is so huge as to dwarf the payoffs of 1000 smaller experiments. There is no evidence of this. 2. There is no world in which this applies to particle physics at this point
344.
by naasking 9mo ago
Wakefield accelerators would be orders of magnitude smaller and higher energy than radio frequency colliders like the FCC. If it has to go to particle physics, this money would be better spent on that research.
345.
by naasking 9mo ago
What a waste. So many more science experiments with better expected ROI could be funded for the money needed for the FCC, and we're not even expecting any significant new insights from it.
346.
by naasking 9mo ago
That's not an answer to the specific question asked. Not all types of fundamental research have the same potential for material benefits, or the same cost.
347.
by naasking 9mo ago
That's really interesting. I ran this scenario through GPT-5.1 and the reasoning it gave made sense, which essentially boils down to: in tools like Claude Code, Gemini Codex, and other “agentic coding” modes, the model isn’t just gener
348.
by naasking 9mo ago
IU was correctly used everywhere else in the article except that one place with the mistak, so the LLM didn't hallucinate a correction, it correctly summarized what the bulk of the article actually said.
349.
by naasking 9mo ago
> Chemistry trumps psychology. Good enough chemistry enables cognitive treatments. But to fix the wrong chemistry you need chemistry. It's not at all clear that chemistry is the root issue. The brain is a synaptic graph that does so
350.
by naasking 9mo ago
> I wouldn't have believed it a few tears ago if you told me the industry would one day, in lockstep, decide that shipping more tech-debt is awesome. It's not debt if you never have to pay it back. If a model can regenerate a w
351.
by naasking 9mo ago
> That's not a like for like comparison, and that site is bonkers in that it's asking models to make nonsense up. LLMs are designed to make things up, it's literally built into the architecture that it should be able synth
352.
by naasking 9mo ago
> Ie: Grok generating CP makes it completely unusable in a professional setting Do you mean it's unusable if you're passing user-provided prompts to Grok, or do you mean you can't even use Grok to let company employees wri
353.
by naasking 9mo ago
Did we all forget about the censorship around "misinformation" during COVID and "stolen elections" already?
354.
by naasking 9mo ago
> Instead of judging, we just note that somehow we accidentally created a replicator, and replicators are going to replicate until something makes them stop. I think this was kind of known for a long time, and pithily described in the we
355.
by naasking 9mo ago
AIs are much better and hallucinate much less when they are given focused tasks, eg. instead of asking AI to write a background on complete with citations, ask the AI specifically to generate a list of citations relevant to X, programmatica
356.
by naasking 9mo ago
Actually, cities often regulate gardens: https://www.savingadvice.com/articles/2025/07/07/10160132_th...
357.
by naasking 9mo ago
> People who love using AI to create software are loving it because they don’t value the act of creating & understanding the software. Most people have a car just to get around, it's true, and some people love tinkering with car
358.
by naasking 9mo ago
That depends entirely on the specifics. If the worst revolution killed 20 people and the best led to the Enlightenment, scientific progress, vaccines, etc., you absolutely can judge your claim's merits, even if there were 10,000 exampl
359.
by naasking 9mo ago
> Encourages cognitive offloading, leading to skill atrophy Speculative, we don't have that evidence yet. The evidence we do have is in students who have not yet developed skills and maturity, and who also have a lot of other conf
360.
by naasking 9mo ago
It could be if it actually lets us calibrate our credence of your original claim that most revolutions have resulted in a lot of death for little benefit. If the worst examples are much worse than the best examples, or vice versa, then we c

More ›