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

531–540 of 626 posts

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

#531

Post by Ydarbleoj »

bpodgursky wrote: post #43
kouteiheika wrote: post #39

Earlier quoted context omitted.

> I don't understand all this spite about "rich friends" Okay, here's a challenge: I assume you're not a rich and powerful entity, so try to gain access to Mythos. I'll wait. > I mean what honestly are you thinking Anthropic can do to give you better cyber tools? Their frontier model was literally nuked by the feds for a month for doing it. Well, first I'd suggest they stop with the constant fear mongering. Here's my…

This is a lot of words to say "you're right, Anthropic does not have any legal way to release frontier cyber capabilities to the public"

I still think it's more about the cost of compute than security.

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

#532

Post by mxmilkiib »

Garpagan wrote: post #521
weego wrote: post #489

Earlier quoted context omitted.

what is a harness? The comments below are mixing IDE/ADE but other suggestions are purely terminal things and I don't get what their value is over just a terminal. Is a harness like a loop where it's just a vague thing that everyone nods about but everyone is nodding at something different?

Harness is a set of artisanal prompts.

isn't that more what a skill or mpc are?

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

#533

Post by codethief »

somenameforme wrote: post #90
MangoCoffee wrote: post #53

Earlier quoted context omitted.

OpenAI and Anthropic are both seeking trillion IPOs, while Chinese labs are pumping out open-weight models that are free for US providers to host and monetize. These Chinese models cost less of US SOTA models to run, even if they are less capable. Providers can just run them, offer cheap tokens, and pocket the margin. I just don't see how you justify a trillion valuation for US AI labs when the underlying models are…

Another interesting potential market here will be 'LLM in a box'. All the hardware and other tooling in a prebuilt, but modular, package ready to go. Pay one up-front cost, get a system running [whatever open LLM] with a token rate of [x], optionally configured to be immediately ready for distributed usage. Basically the opposite of cloud stuff: no rent, no dependency, 100% guaranteed uptime, guaranteed security/priv…

Exactly! As I've argued here on HN before, such an "LLM in a box" might end up being serviced/upgraded once or twice a year by a company very similar to the one servicing the coffee machine at the office. In contrast to databases, storage, etc. it doesn't matter much if the box breaks at some point – they'll just come by and replace it with a new one – and there's barely any software on the box to speak of, at least none that requires continuous development and feature upgrades, beyond rolling out security patches. This makes the business case drastically different from cloud and SaaS offerings, where most of the moat is in the software and the state maintenance (and the vendor lock-in of course). The LLM in a box is destined to become a commodity.

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

#534

Post by ccgreg »

alightsoul wrote: post #85
Gecko4072 wrote: post #15

People familiar with the topic, how will models continue to get better? Post training it seems? Labs have already used up internet-scale data, so are there any limits to architecture improvements and post training or can we expect this trend to continue? ByteDance is training a 10T-parameter model. Here, GLM 5.3 outperforms models 3-4x its size of roughly 700B, so parameter count doesn’t seem to be a direct correlati…

GitHub dumps are about 115 terabytes. The common crawl is in the petabyte range uncompressed for every year. Apparently there are dumps of Reddit too in spite of their efforts to ban bots and it's not solely due to the use of residential proxies. For a 1:20 parameter to token ratio, you can still train up to 10 trillion parameters so 10T parameters times 20 is about 200 trillion tokens. Then each token is 4 bytes so…

Common Crawl is about 1 petabyte per year, compressed. Uncompressed is 4X.

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

#535

Post by h8hawk »

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 capture. Otherwise, the widespread availability of good tooling for security analysis would eliminate more of these cyber threats, rather than gatekeeping them for a few private companies.

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

#537

Post by xscott »

cosmez wrote: post #530
xscott wrote: post #91

Earlier quoted context omitted.

Probably not what you're after, but I've considered having a separate small mm-model act as a seeing-eye dog for the bigger more capable one.

But how do you prompt this smaller model to give back information? I've tried that in the past but didn't go well. What I do is to send written handoff files between models to pass context around, but only had good results with big vision models as well.

So many possibilities for how you could glue it all together. However, when I send Gemma 4 12B in llama.cpp an image with no accompanying text, it assumes I want a description and gives me one.

I just tried with an audio file, and it transcribed the lyrics as I hoped. Then it made a bunch of suggestions about what do next, which I wasn't after. I could probably fix that by sending some text to narrow the scope.

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

#538

Post by theappsecguy »

standeven wrote: post #473
Gigachad wrote: post #124

Earlier quoted context omitted.

This is going to be catastrophic. Whether AI works or is useful or not isn’t even the question anymore. It can fulfil every promise Sam Altman has been making and will still make no financial sense to justify these valuations.

The US will just do what they did with Chinese EVs: ban the superior technology to protect US companies.

a lot harder to ban software than hardware the size of EVs

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

#539

Post by flanked-evergl »

jauntywundrkind wrote: post #487
flanked-evergl wrote: post #465

Earlier quoted context omitted.

The v2 branch of OpenCode has not been touched for months, if it's beeing developed then I don't know where.

this is very easy to check. i promise i'm not tricking you and just uploaded this. https://github.com/anomalyco/opencode/tree/v2 this is the integration branch for https://opencode.ai/v2 . it has been for months. it's where the Effect-based refactor has been landing.

sorry, I was mistaken! I looked at https://github.com/anomalyco/opencode/tree/2.0

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

#540

Post by redeeman »

badcafe23423435 wrote: post #480
KronisLV wrote: post #342

Earlier quoted context omitted.

What the person above is suggesting: * https://pi.dev/ * https://omp.sh/ (no personal opinions of either, links might be useful) I think that OpenCode is nice, their CLI version is enjoyable and their desktop/web version is okay : * https://opencode.ai/ I also quite like driving OpenCode through something like Kepler / Paseo and tools like that (with those I can still use my Anthropic Condition by Claude Code being t…

No one in their right mind would install software using `curl | bash`

its a good way to check if people are insane though. would be a cool tactic for new hire evaluation, monitor them setting up dev environment. do the curl | bash, and its instafail
Post reply on HN