Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

The Closing of the Frontier

tanyaverma.sh

Post reply on HN
Print view View original

131–140 of 142 posts

Re: The Closing of the Frontier

#131

Post by MisterTea »

dreamcompiler wrote: post #109
bombcar wrote: post #86

Earlier quoted context omitted.

There’s a drain clog clearer sold in a jug like all the rest. But they wrap the jug in a thick clear bag. The implication is clear - this stuff is so powerful it’s extra dangerous. It’s the same stuff inside as all the others.

No. That stuff is sulfuric acid and it needs secondary containment during shipping, which the ordinary drain cleaners (usually diluted sodium hydroxide) generally do not.

This is correct. That stuff is also horrible for cast iron pipes which are code mandatory in many cities such as NYC. Doesn't stop stores from selling it or stupid people from using it in their cast iron pipes.

Re: The Closing of the Frontier

#132

Post by adrian_b »

laurentiurad wrote: post #49

My bet that this is just a model with a slimmer safety alignment which yields results in 100 agent running in parallel configuration.

They have said that their harness that is used for finding bugs is running many times Mythos for each source file, with increasingly specific prompts, depending on the previous results, and all the independent runs are done in parallel.

Re: The Closing of the Frontier

#133

Post by altmanaltman »

contraposit wrote: post #83
andyfilms1 wrote: post #36

Earlier quoted context omitted.

It has always baffled me how quickly, and how voraciously, people started to rely on privately owned AI systems. AI is not something discovered by scientists and plucked out of the ether. It's engineered and controlled, for profit, by corporations which have demographics and KPIs. These companies don't owe you anything, and they make no promises. If you're running a business that deeply relies on AI, you might as wel…

I can't run my business without electricity. Yet we don't fear of its access being revoked. Sam makes the comparison of intelligence to electricity a lot. So we are on the path to these systems becoming utilities.

Comparing AI to electricity focusing on just one particular aspect (hey its like fuel guys!!) while completely ignoring all the structural difference between actual energy industries and big tech is really stupid.

Re: The Closing of the Frontier

#134

Post by htrp »

bcatanzaro wrote: post #129
htrp wrote: post #117

Earlier quoted context omitted.

how do you justify the compute investment for something like nemotron ? especially if all the labs are willing to pay for those same GPU clusters for inference or training runs?

Nemotron has two reasons to exist, both of which are strategic to NVIDIA. 1. Help NVIDIA design future systems for AI by more deeply understanding what it takes to build AI. 2. Keep the AI ecosystem strong and diverse throughout the world by providing AI infrastructure that many companies can innovate on. This is not a science project, nor is it for the joy of giving something away. Both of these reasons are core to…

Does Nvidia maintain it's own compute hardware expressly for model training? Otherwise, I'm not sure how you keep up with the SOTA model techniques.

Re: The Closing of the Frontier

#135

Post by altmanaltman »

signatoremo wrote: post #111
anematode wrote: post #90

Earlier quoted context omitted.

Electricity is heavily regulated. Is there any evidence that LLMs will be the same?

Was electricity regulated in the first decade of its existence?

Was anything regulated in those times? You could legally buy humans at that time.

But that doesn't mean we live with same standards. Lack of regulations in electricity led to a lot of deaths and disaster which is why it was regulated.

But we dont live in the start of 20th century, we live in 2026 and we must learn from the past instead of helbent on repeating it.

Re: The Closing of the Frontier

#136

Post by adrian_b »

markhahn wrote: post #114

mythos has not been demonstrated doing anything dramatically different than other models. so as other comments say: very premature. but the basic premise (shared among a lot of ai-doomers and ai-shamers), is that the bigs have somehow raped society (by training on everything available). this needs to be challenged: it implies quite a strong model of IP ownership, which is not what appears in law, or in founding docum…

"Raping society" seems an appropriate term, though a better metaphor is cutting a public forest or mining in public lands, which are other examples of converting public properties into private properties.

Actually I, and apparently many others, would have no problem with the fact that companies like OpenAI or Google have gathered huge amounts of information from all over the world into their training sets.

What bothers me is that I do not have access to the same information. If I would try to run a bot, it would be blocked immediately by all sites. If I would copy pirated books or movies, that is supposed to be punishable legally.

None of what is forbidden for me was forbidden for the big companies. What I want is not that they should be punished, but I want for myself and for any other people the same rights, i.e. access to the same training sets.

For now, I must be grateful that a part of the hoarded information is available for the public in a non-deterministic manner in the existing open-weights models.

This is much better than nothing, but I would prefer access to the training sets, even if in that case the AI companies would keep for themselves any trained models. There are many tricks that they have used during training, but by far the input training data is much more important, since anyone can discover better training algorithms.

What I find unacceptable is that now they consider as their private property what they have mined from public lands.

Re: The Closing of the Frontier

#137

Post by blastslot »

adrian_b wrote: post #130
skybrian wrote: post #45

It's long been conventional wisdom that you shouldn't write your own crypto libraries - leave that to experts. But excellent open source libraries are available, which do get reviewed by experts. And if you're willing to study, maybe you can learn enough about cryptography to become one of the experts? I'm wondering what other security-sensitive software that might become true of in the era of Mythos-or-better AI's?…

Someone must have written the existing crypto libraries, so obviously they have not followed the conventional wisdom. The advice that you should not write crypto libraries must be taken very seriously, but you must understand that its meaning is just that this is one of those things that is harder than it seems, so you must not try to do it before being thoroughly prepared for this, because there are no shortcuts, an…

[dead]

Re: The Closing of the Frontier

#138

Post by tim-projects »

avaer wrote: post #69
tim-projects wrote: post #48

Anthropic should provide a specific service where they attack a businesses infrastructure using this frontier model and then issue a report of all vulnerabilities found. I could imagine it would be quite lucrative. Much better than hiding it away where it can't help anyone.

Hypothetically... I'm joey joe bob who happens to be maintainer of a top 10 npm package. My wife got cancer so I need money FAST. Unrelated: can you mythos my lib? 1 month later: whoopsie my lib got hacked and the hackers stole a bunch of stuff. Sorry guys. It just wouldn't be good PR. And this is the best case scenario.

Your scenario has little to do with what I said.

Re: The Closing of the Frontier

#139

Post by vector_spaces »

hn_throwaway_99 wrote: post #106
vector_spaces wrote: post #89

Earlier quoted context omitted.

> pithiness is more important than ever I apologize for getting stuck on your parenthetical but while pithiness is a fine aspiration in a North American business setting, pithy reads generally can't exist without more detailed and nuanced long-form analyses, and the latter face a more dire existential threat. You are right that pithy [writing] is an important skill, as are slow and deliberative reading and writing of…

To give a clearer example of what I was talking about, look at the linked article from my comment above. It is much longer than the original article from this post, but it (a) starts with a TL;DR, so it gives me a summary that lets me know if I want to read it in the first place, (b) combines detailed original research with analysis and opinion, and, IMO (c) continually adds new information/insights so it builds on i…

OK, that's fair -- it sounds to me though that it's less about pithiness, or that this word might be too narrow. To me it sounds like you are talking about accessibility, in the sense that you'd like more care from writers in structuring their writing with consideration for legibility to the reader. In that I am with you

Re: The Closing of the Frontier

#140

Post by bcatanzaro »

htrp wrote: post #134
bcatanzaro wrote: post #129

Earlier quoted context omitted.

Nemotron has two reasons to exist, both of which are strategic to NVIDIA. 1. Help NVIDIA design future systems for AI by more deeply understanding what it takes to build AI. 2. Keep the AI ecosystem strong and diverse throughout the world by providing AI infrastructure that many companies can innovate on. This is not a science project, nor is it for the joy of giving something away. Both of these reasons are core to…

Does Nvidia maintain it's own compute hardware expressly for model training? Otherwise, I'm not sure how you keep up with the SOTA model techniques.

Yes
Post reply on HN