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

Qwen3.8-Max: A New Bar for Coding and Cowork

qwen.ai

Post reply on HN
Print view View original

561–570 of 652 posts

Re: Qwen3.8-Max: A New Bar for Coding and Cowork

#561

Post by desterothx »

mbrumlow wrote: post #553
delusional wrote: post #540

Earlier quoted context omitted.

What amazing windows drivers have you sold?

Idk about sold. But it’s loaded on all the windows machines in a fairly big company that solves a real world problem. It allows us to apply custom ACLs to AI agents and the child process spawned by AI agents. Giving us the ability to control what files an AI agent can read or write to, while still being in the calling users context. It allows us to force all ai derived processes to use a transparent MITM proxy so we…

so you decided to vibe code a sandbox because windows sandboxing is lacking, instead of moving to linux where sandboxing is robust?

Re: Qwen3.8-Max: A New Bar for Coding and Cowork

#562

Post by blitzar »

throw-the-towel wrote: post #560
blitzar wrote: post #531

Earlier quoted context omitted.

> they need to see/learn by themselves The cleanup pay is mighty generous as well.

Has anyone gotten a position as a post-vibe janitor yet?

You got problems there too, but I don't do shitters. That's Engineering.

Re: Qwen3.8-Max: A New Bar for Coding and Cowork

#563

Post by ilaksh »

zdragnar wrote: post #559
ilaksh wrote: post #546

Earlier quoted context omitted.

That's a popular type of AI cope. First, I am aware they have been able to write code for some time. I made my first LLM coding agent experiment the day after ChatGPT first blew up in November 2022. I haven't been a "code monkey" in the last say 15 or 20 years of my programming. I would argue also that code money was never really a thing, it was just an excuse that pretentious overpaid developers used for the existen…

Two points of disagreement: "Code monkey" absolutely was and still is a thing. I was hired by Big Name Co originally to be staff but a reorg had me do typical senior position duties for around a year, and it was hell. Every decision was made over our heads, feedback was unwelcome because the feature had already gone through design and gotten layers of sign-offs from different areas of the org, etc. My sole purpose wa…

Those are good points, but that mostly comes down to lack of organizational status or office politics etc. which is independent of capability level and largely orthogonal to whether it's a human programmer or an AI agent. Unless you are redefining "software engineer" to be more like a political operator. But that goes for any human group and any job. And anyway after a certain level, not providing the senior engineers and or project managers with sufficient political sway to realize projects goals is a management failure. Unless again you are defining software engineer as someone who is a master social manipulator in the context of their job.

Re: Qwen3.8-Max: A New Bar for Coding and Cowork

#564

Post by mbrumlow »

desterothx wrote: post #561
mbrumlow wrote: post #553

Earlier quoted context omitted.

Idk about sold. But it’s loaded on all the windows machines in a fairly big company that solves a real world problem. It allows us to apply custom ACLs to AI agents and the child process spawned by AI agents. Giving us the ability to control what files an AI agent can read or write to, while still being in the calling users context. It allows us to force all ai derived processes to use a transparent MITM proxy so we…

so you decided to vibe code a sandbox because windows sandboxing is lacking, instead of moving to linux where sandboxing is robust?

Some products and companies require windows.

It might please you to know that this driver was vibe coded from a linux vm based sandbox.(that was also vibe coded)

Re: Qwen3.8-Max: A New Bar for Coding and Cowork

#565

Post by WarmWash »

delusional wrote: post #540
mbrumlow wrote: post #523

Earlier quoted context omitted.

What does the benchmark even mean when people are using AI to make real world things that solve real world problems? I see people, and my self making amazing things with AI and fixing old projects and having real world impact at the fraction of the cost it would take me to hire people, or hours spent on my own coding. I have built tools and systems with AI that have allowed me to build windows drivers, android apps,…

What amazing windows drivers have you sold?

A better question is which products/services have you avoided purchasing.

Re: Qwen3.8-Max: A New Bar for Coding and Cowork

#566

Post by nearbuy »

whage wrote: post #322
ucha wrote: post #264

Earlier quoted context omitted.

That's not how LLMs work. If you're talking about number of parameters, you wouldn't be able to reduce the size much by "removing" support for other languages.

Can you please explain why?

Only a tiny, tiny fraction of the parameters are encoding information that's specific to a particular programming language. Even if you could remove those without degrading performance, it would have a negligible effect on the model size.

Re: Qwen3.8-Max: A New Bar for Coding and Cowork

#567

Post by burnte »

blitzar wrote: post #531
gtirloni wrote: post #490

Earlier quoted context omitted.

> Anyway the point is these models are just about capable of doing the entire job of analyzing a small business and building out all the agents and iterating on them with the business owner. If you search for LLM benchmarks focused on real-world tasks, you'll quickly learn this ain't the case. No point in telling business owners about that though, they need to see/learn by themselves.

> they need to see/learn by themselves The cleanup pay is mighty generous as well.

I just finished two solid months of my life cleaning up a vibe coded app.

Re: Qwen3.8-Max: A New Bar for Coding and Cowork

#568

Post by dgellow »

ilaksh wrote: post #554
benatkin wrote: post #516

Earlier quoted context omitted.

To say that you're building agents, is based on a way of viewing things that isn't at all pervasive. Some might say you're building a customer support chatbot. To talk of building agents as a common activity makes sense if you use LangChain I guess, where its title is "Open Source AI Agent Framework | Build Agents Faster". But for instance, YC just released qm, and in its README it talks about customizing the agent,…

Not sure what you are trying to say but thanks for reminding me about qm, it has some cool features. My MindRoot framework had some of qm's main features awhile ago though. I do try to use it to build solutions by customizing rather than from scratch when possible. But a lot of clients or potential clients don't like that idea even though it's MIT. They want some that has clearly been invented by them as a new progra…

Why are you looking for projects on upwork if you know they don’t pay well and that you don’t have leverage? That sounds like a pretty bad place where to find contracting projects

Re: Qwen3.8-Max: A New Bar for Coding and Cowork

#569

Post by dzonga »

tomjen3 wrote: post #522
ilaksh wrote: post #471

As someone who is searching for a new programming contract right now, reading all of the incredible abilities here is pretty intimidating. Especially since I get almost all of my projects from Upwork which is an outsourcing site. I believe I am competing directly with these frontier models in some circumstances. Like there are a ton of programmers who previously would be outsourcing work to that site, but now they as…

As someone else who is an IT consultant. You need to run away from any client who thinks your primary purpose is to write code. You must run away from the business if YOU think your primary purpose is to write code. You should be able to write code and do it well, But the AIs, and I'm not even talking about the ones on the frontier, have been able to write code faster than I have for quite some time. What you need to…

bingo. to give an analogy in terms of law firms - you can either be a partner who brings in new business or the associate who does the grunt work of reviewing/writing contracts.

guess one which is valuable ?

Re: Qwen3.8-Max: A New Bar for Coding and Cowork

#570

Post by zdragnar »

ilaksh wrote: post #563
zdragnar wrote: post #559

Earlier quoted context omitted.

Two points of disagreement: "Code monkey" absolutely was and still is a thing. I was hired by Big Name Co originally to be staff but a reorg had me do typical senior position duties for around a year, and it was hell. Every decision was made over our heads, feedback was unwelcome because the feature had already gone through design and gotten layers of sign-offs from different areas of the org, etc. My sole purpose wa…

Those are good points, but that mostly comes down to lack of organizational status or office politics etc. which is independent of capability level and largely orthogonal to whether it's a human programmer or an AI agent. Unless you are redefining "software engineer" to be more like a political operator. But that goes for any human group and any job. And anyway after a certain level, not providing the senior engineer…

A big part of an engineering performing requirements analysis is owning the technical understanding of the requirements and identifying areas where the requirements are not technically feasible, can be done with less expense with adjustments, or can be better achieved with a different solution.

AI will help you draft the communication of that knowledge, but it won't help you actually deliver the knowledge or ensure that the knowledge was appropriately received. As you say, it is a fundamentally human problem, which is why I maintain that writing code is easier.

Post reply on HN