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

GLM-5.3: Frontier coding with emergent cyber capabilities

z.ai

Post reply on HN
Print view View original

601–610 of 626 posts

Re: GLM-5.3: Frontier coding with emergent cyber capabilities

#601

Post by bigbadfeline »

andxor wrote: post #523
justapassenger wrote: post #419

Earlier quoted context omitted.

It’s not that frontier providers won’t keep on making good/leading models. It’s whether you absolutely need the latest capabilities (at the cost of very high prices, sending your data to them, and being totally at the whim of 2 companies, that can shut you off anytime for any reason). With how good LLMs are already, there’s tons of tasks where not being at the absolute bleeding edge doesn’t matter, especially when yo…

There's no shortage of extremely valuable problems to solve.

> There's no shortage of extremely valuable problems to solve.

And there's no shortage of cheap models that are perfectly capable of solving them.

Re: GLM-5.3: Frontier coding with emergent cyber capabilities

#602

Post by inigyou »

eru wrote: post #584
inigyou wrote: post #577

Earlier quoted context omitted.

The computer price crisis is also the fault of central banks, since they printed the money and gave it to the AI companies to buy everything with.

Not really. Central banks don't really control relative prices.

They control the allocation of new money.

Re: GLM-5.3: Frontier coding with emergent cyber capabilities

#603

Post by bigbadfeline »

WarmWash wrote: post #383
vmware508 wrote: post #106

Apple will release M7 MacBook Pros / Mac Minis next year, and they will be able to run free LLMs locally at native speed. All software developer notebooks will be replaced to run local models, saving a lot by cancelling Claude Code subscriptions. Developers win. Apple stocks will be rocketing. Everything else will go down. You're welcome.

I'm still waiting for Linux to topple Windows

The processors running the Linux kernel outnumber those running Windows 4:1.

Re: GLM-5.3: Frontier coding with emergent cyber capabilities

#605

Post by alightsoul »

smurf9852 wrote: post #292

" a judge agent then attempts each task to verify that it is actually solvable " I understand you need to verify the goal is achievable. But if the judge agent has the same goal as the training agent (solve), and both are of the same model, then aren't the judge and the training agent doing the exact same thing? What is the point then? Can someone explain this to me.

It works because models are not deterministic so you can prompt two instances of the same model to be adversaries of each other

Re: GLM-5.3: Frontier coding with emergent cyber capabilities

#606

Post by cubefox »

h8hawk wrote: post #535
cubefox wrote: post #84

> Open Source: We will release the weights in two weeks after launch, once safety evaluation and hardening are complete. What safety evaluation? What safety hardening? They already evaluated it and found it to be highly capable at exploiting security vulnerabilities. So we know it is not "safe", and they don't seem to plan to do anything against it. What could be more dangerous than hacking? Biological weapons resear…

Are you against open-source models? Data and content related to "biological weapons" already exist on the internet, in books, etc. The real issue is access to facilities and tools. There are models that help researchers, but they are not LLMs, rather they are models trained specifically on biological data (like AlphaFold). Cybersecurity is basically used like a dog whistle pioneered by Anthropic to achieve regulatory…

> Data and content related to "biological weapons" already exist on the internet, in books, etc. The real issue is access to facilities and tools.

No, I think tools are easy to come by (unlike in nuclear research), the real issue is the know-how to create biological weapons, which you can't easily get out of books, but much more easily out of an amoral LLM.

Re: GLM-5.3: Frontier coding with emergent cyber capabilities

#607

Post by Sabinus »

ShinyLeftPad wrote: post #593
aliljet wrote: post #5

This is absolutely still shy of Sol and Fable, but only just by a hair. Ridiculous results. There's still not a compelling economic reason to drop OpenAI courtesy of the ludicrous reset addiction that's taken place, but it feels like we're on the precipice. How are you all toying with running this kind of thing in a mega quantized way locally? Two weeks out from released weights, but this is still just GLM 5.2 with p…

What is reset addiction?

OpenAI, or specifically one guy on Twitter, seems to be so regularly resetting quotas that it's becoming the new normal and it'll suck when it stops happening.

Re: GLM-5.3: Frontier coding with emergent cyber capabilities

#608

Post by h8hawk »

cubefox wrote: post #606
h8hawk wrote: post #535

Earlier quoted context omitted.

Are you against open-source models? Data and content related to "biological weapons" already exist on the internet, in books, etc. The real issue is access to facilities and tools. There are models that help researchers, but they are not LLMs, rather they are models trained specifically on biological data (like AlphaFold). Cybersecurity is basically used like a dog whistle pioneered by Anthropic to achieve regulatory…

> Data and content related to "biological weapons" already exist on the internet, in books, etc. The real issue is access to facilities and tools. No, I think tools are easy to come by (unlike in nuclear research), the real issue is the know-how to create biological weapons, which you can't easily get out of books, but much more easily out of an amoral LLM.

This is just wrong. Standard textbooks like "Bioprocess Engineering Principles" [1] already contain step by step industrial engineering protocols and formulas. LLMs only summarize what print literature has documented for decades. The real bottleneck has always been hands on physical execution, not access to text. (This is not the case with actual models trained on biological data, which are not LLMs)

> amoral LLM

Are you a Lesswrong member? What makes Anthropic moral when their priority is automating people out of jobs instead of developing better medicine or advancing healthcare?

[1] https://www.sciencedirect.com/book/monograph/9780122208515/b...

Re: GLM-5.3: Frontier coding with emergent cyber capabilities

#609

Post by simonjgreen »

kouteiheika wrote: post #44
simonjgreen wrote: post #33

Earlier quoted context omitted.

We applied for the cybersecurity approval via the form and got approval back in less than an hour. Have you… tried?

Have you tried to use Fable for anything even remotely security related, when the refusals kick in as soon as you even fart in the vague direction of anything security or biology-adjacent?

Our automated CVE testing harness is now powered by Claude after we got the approval. (For interest, we’re using VVAH https://github.com/visa/visa-vulnerability-agentic-harness )

Re: GLM-5.3: Frontier coding with emergent cyber capabilities

#610

Post by SXX »

embedding-shape wrote: post #592
SXX wrote: post #581

Earlier quoted context omitted.

Chemical weapons are not used not because some agreements - they just messy and only good for killing civilians. Also contaminate area and might also kill your own personnel. If you look at all other banned weapons they are all used against Ukraine by Russia and nobody gives a damn.

> Chemical weapons are not used not because some agreements We've literally destroyed countless of supplies of chemical weapons because of agreements about not using them, because we all agree they're absolutely horrible: > The OPCW administers the terms of the CWC to 192 signatories, which represents 98% of the global population. As of June 2016, 66,368 of 72,525 metric tonnes, (92% of chemical weapon stockpiles), h…

Yes chemical weapons has been destroyed and it's a good thing, but all other efficient and practical weapons weren't successfully banned.

Even EU countries that signed agreements against anti-infantry mines exiting them because how efficient they are at slowing down invasion.

Post reply on HN