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

271–280 of 626 posts

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

#271

Post by cyanydeez »

andsoitis wrote: post #240
grey-area wrote: post #233

Earlier quoted context omitted.

Circular investment deals and investment deals at valuations which have no possible justification.

Let me ask it differently. You state the companies and their investors are colluding. Who are they colluding against ?

its interesting, as it's typically the banks and against the public at large because the goal is to jimmy up valuations to justify IPOs then sell on opening; just like spacex.

It's what enron was doing; it's what most of crypto's offshoots were doing.

Sure you can blame the marks of the grift and say "well the public should know they're faking all this cash flow expectation".

It seems like you're either driving the grift economy or part of the collusion.

It's similar to how a cult operates, so I'll be frank: your skepticism seems biased.

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

#272

Post by mbil »

andxor wrote: post #127
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…

Fable finished training 6+ months ago. At this point, Anthropic only needs to release models to the public when the competition forces them to. OpenAI also has a better model (Astra) that they haven't released yet.

Yes it seems like the thread is discounting that frontier providers are likely already baking new, stronger models. I agree that GLM and its ilk are quite good, but having used them I’m not convinced they’re on par with eg Opus in terms of things like tool calling. And they’re fast but less capable so I spend about the same amount of time with them, just with more hand holding. Maybe this is a harness limitation. I know on paper they seem comparable but anecdotally and qualitatively they’re not as useful as the frontiers’, so maybe there’s some truth to benchmaxing claims. For some workloads the distilled models may be good enough, and I suspect at some point there will be diminishing returns to spending a premium on frontier models, but I don’t think we’re there yet. That said I’m continuing to try them.

The question is whether this steals enough marketshare from frontier providers that they don’t have the capital to train the next model iteration. The open models are going to push down the unit price of an intelligence-token, but there will still be a market for a smarter bot. And as intelligence gets cheaper, the demand for it will rise (see Hank Green’s Jevons Paradox video). Not to mention there’s all kinds of other directions to go at the frontier (world models, robotics, video gen, etc).

Another thing, and this is pure speculation, but if the Chinese model providers already discovered the decrypting COT trick and leveraged it to do RL training, and assuming frontiers plug that hole, then maybe future distillation will be harder.

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

#273

Post by Havoc »

unrvl22 wrote: post #107
tjwebbnorfolk wrote: post #40

Earlier quoted context omitted.

does this suggest 5.3 is the same # of parameters as 5.2?

which is the bigger headline that people don't realize. this is 744b and its head to head with Kimi K3 (2.8T), smashes DS v4 pro (1.5T). even Opus and Sol are rumored to be 1.5T+ this is half the size!

You do need to compare active parameter too though. The total size isn’t a reliable indicator anymore

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

#274

Post by onlyrealcuzzo »

alienbaby wrote: post #269

One htought I had; if The chinese allow unfettered access to cyber capabilties while th US does it's best to neuter it's model releases, from China's point of view they have the US all tied up in knots dealing with problems they don't give people the tools to solve. China giggles as it watches the US under threat from people using it's models. The US is restricting citizens from owning this particular kind of weapon,…

I suspect Anthropic wanted the US gov to ban Mythos for marketing.

If it turns out to be bad for them, the US gov will likely suddenly unban models.

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

#275

Post by zozbot234 »

numpad0 wrote: post #234
somenameforme wrote: post #90

Earlier quoted context omitted.

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…

What makes that kinda complicated is that multi-user throughput of LLMs scale well but single-user performance often stays constant at low ends. If you could saturate e.g. 16 concurrent session-month of demand, you can just go buy 16 of 32GB GPUs and start charging monthly for inference. That could work if you had e.g. over thousand total employees with hundreds of devs eager to trying it out, but only if the company…

You're talking about multi-session vs. single-session throughput. A single user can easily leverage multiple sessions via e.g. subagent swarms, especially on a lower-end setup where any single session is going to be quite slow. Saturating utilization during off-hours is harder but potentially quite feasible by assigning lower priority, unattended tasks/inference loops.

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

#276

Post by andsoitis »

cyanydeez wrote: post #271
andsoitis wrote: post #240

Earlier quoted context omitted.

Let me ask it differently. You state the companies and their investors are colluding. Who are they colluding against ?

its interesting, as it's typically the banks and against the public at large because the goal is to jimmy up valuations to justify IPOs then sell on opening; just like spacex. It's what enron was doing; it's what most of crypto's offshoots were doing. Sure you can blame the marks of the grift and say "well the public should know they're faking all this cash flow expectation". It seems like you're either driving the g…

> It's what enron was doing;

Enron hid billions of dollars in debt and fake profits.

Is this what you think is happening here?

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

#277

Post by alienbaby »

dhx wrote: post #161
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.

I take it from [1] (transcript of recent DeepSeek CEO discussion with investors) that DeepSeek would disagree on the immediate catastrophic impact to the likes of OpenAI or Anthropic. The reason is even though technology parity mostly exists, only OpenAI, Anthropic et al have the inference capacity to gain market share and generate revenue. Chinese vendors don't have the chips needed to scale up inference and gain ma…

is that releveant if people can host their own models? that activity still undermines the valuation / diminishes the US companies 'moat' ?

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

#278

Post by orbital-decay »

budsniffer952 wrote: post #256
orbital-decay wrote: post #199

Earlier quoted context omitted.

Why would you even believe the opposite? US spooks have been amassing vulnerabilities and relying on them for decades, they literally pioneered it in the 90's if not earlier. Everyone does it now but the US is the biggest of them all. Surely this devalues a lot of what they did. Moreover, the way the US government handled new capabilities, and OpenAI's training policy (they are in bed with the government) just scream…

[flagged]

Good thing I said nothing of that (especially nothing about China). Reread it again to understand you built an incredible strawman and ignored my last sentence.

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

#279

Post by cyanydeez »

wren6991 wrote: post #139
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…

The thing that blows me away is it does this at one quarter the total parameter count of K3 (and 40% active parameter count). There's plenty of room at the bottom. > How are you all toying with running this kind of thing in a mega quantized way locally? Sure, let me answer that in excessive detail. I briefly tried running the UD IQ3_S quant of GLM-5.2, which is 288 GiB of weights (301 GB). Setup was: llama.cpp, 1x NV…

I'm hoping colibri can start pulling in specifically designed models for the heirarchy of decoding. It seems like we should be able to get smarter MoE models that can do the work.

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

#280

Post by Havoc »

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.

That’s not how that works. The hosted models don’t stay still in size and capability while Apple advances. Both will advance their frontier and there will still be a gap and developers will still prefer the stronger option.
Post reply on HN