bitexploder

If you are okay with waiting use GLM 5.3 max. It costs more but still cheap. It is slow, but a very strong worker. Still dollars per day (at most) with heavy concurrent agent running. I load up planning and tasks in Opus or Sol, and just have glm flash workers go to town every night. My project has never advanced more smoothly.

Which versions of flash and at what thinking levels? Which chinese flash models and at what thinking levels? What tasks? What completion rates? How was quality evaluated?

Yep. It really isn’t nefarious or abusive. I keep mine slowed down to not scan often to be considerate. Even private trackers can require it now.

For what it’s worth, I am very happy with Jellyfin and the *arr suite. It took a bit of agent prodding to get them all playing nicely together and bypassing Cloudflare CAPTCHAs. However, it's pretty sweet when you get it all working.

It does seem like for new code that might help. There's some really good logic and wisdom in it, but it has to be applied very contextually to the exact problem you are trying to solve. If an agent is navigating a complex codebase, this could definitely send them off on a refactoring rabbit hole. However, if you have them writing some new code, it could prevent their tendency to yak shave and write new things. So I can see some situational uses for this, but it could get out of hand as well.

Sol is my current favorite model to interact with. So much less BS than Opus 5. Fable 5.1 is okay as is Fable 5 but it has Opus like tendencies. Sol is very good at following instructions and remembering them for a session.

Amusingly, as an autonomous coding agent, I kind of like Opus 5. But I have to bound it on tasks or it just goes off the rails. But I'm bounded tasks, it is genuinely solid. It's kind of like the new Sonnet 5. Right now my favorite model to interact with on the frontier side is Sol 5.6 so I have been using that as my coordinator. Flash 3.8 is my other favorite just because it is so fast and I use it a lot at work and know its quirks.

I still hold the line on interviewing. Maybe some don’t. I am sure it is true. My team us too small with too much responsibility to tolerate mediocrity to any real extent

In modern America the answer to that question is often resoundingly yes. Not just hypothetical.

I believe. I run it on my mac M5 pro at like 30t/s with some RAGs and let it work on stuff overnight and it's great. It isn't the same as the big models where things can be more unbounded, but if local models keep progressing there is a universe where a 200-300B model is all most of us will need to stay out of the big tech moats.

I have a few attention and finish mechanisms in my prompts. I have been using it for a week and a half and with some prompt taming it is great. (I have early access to the models cause I work at the place that makes the model). None of my attempts to ever tame Opus 5 have worked.

Fable is okay, just slower, eats tokens and not any better at coding tasks. Maybe a little better, but not better enough. It's a lot faster to have a cheap and fast flash agent / sonnet do the implementation work with Fable tagging cleanup and divergence from spec and goals.

Flash 3.8 is genuinely my favorite all around model right now. And yeah Opus 4.6 was the last Opus model I liked. 4.8 is tolerable. Opus 5 is a terrorist. It just can't follow an instruction to save its life and regresses rapidly. Sol at least stays on track so I have to smack it's hand way less often. I am biased, but Flash 3.8 and 3.7 are the first Gemini models I just recommend to others.

Opus 5 is a genuinely infuriating model. I hate it’s behavior.

I feel like a lot happened this week and people are glazing how ridiculously strong Flash 3.8 is right now compared to Fable/Opus/Sol/Astra.

I wish people could see how some of this reads. You are an “amateur” using a model 6-8 weeks behind? Really? Sigh.

Does Claude not allow third party harnesses?

But running that fast… with a local RAG? Yeah, it is a very interesting model. Maybe you don’t need a lot of parameters, just a really big local database :)

The thing I didn’t realize for a while is 27B is rather smart. As many (or more) activated parameters as the flash models of the universe that we know about. It reasons very well. It just doesn’t have a lot of knowledge.

/me points over at Flash 3.8 :)

It is more fun than serious at this point. Don't overthink it :)

It is still going to be better at text work, skills, document review, deep reasoning, architecture review, etc. It is only 6 months old, it isn’t like its world knowledge and software knowledge is really out of date. Use it to churn on harder design problems.

Inefficiency is their business model. They just soak $$ into their system. More overall revenue even if most of it is some rube goldberg machine of apathy.

Models are like programming, language languages. You don’t need to follow them. If you don’t do video stuff don’t worry about it. You could probably digest mode every month or two and be just fine.

I get that. I use agents a lot and LLMs often reason with code. It is valuable. I just think the floor is a lot lower for general reasoning and common tasks like that. And in 6-12 months it won’t matter. Google will publish better models. The temporal distortion of how long a Sol or a Fable has existed is real. No one is suddenly missing out on some giant competitive edge because their model is a few months behind. I feel like it’s all just going to normalize and things other than how well your model can write code will matter more and more in 12 to 24 months.

Not really. 2 years ago that was a pretty normal amount of GPU hardware for a hacker or gamer. It's all relative. They are not accessible to most people yet, but for someone that cares and is a technologist? Likely accessible.

AI / LLM is about more than agentic coding. It is one of the least interesting use cases to me, thinking more broadly. HN may be over-indexed on it.

I work there. I have zero internal knowledge about the model. Opinion my own, etc. I don't think it is worth fighting to win on a month to month time horizon. When you step back and look an inch above this market, Gemini Pro 3.1 as a product was released in February. 6 months. It feels like forever and that Google is behind, but on a 2-3 year horizon? The models are going to stay similar.

Also, look at Flash 3.5 to 3.7. Flash 3.7 is a genuinely decent Sonnet 5 class model. Flash 3.7 is quite efficient too. Also, whatever was spent training 3.5 pro is probably not wasted. However, as a strategy, when I see models like Kimi K3, Fable, Sol. If you discard "because the model sucked" what other alternatives or potential options might exist?

I thought of a quite a few and they are far more compelling and interesting to me.

(Also Gemini models tend to be pretty decent at more than just programming. Enterprise AI use is more than just software eng / programming)

I only like talking to Opus 4.6 in its default form. I have some pretty aggressive prompting strategies ensuring my language guidance rules are front and center and it really helps with Opus 4.8+, something went wrong with those models.

Also, OMP has a solid subagent model and I like it with some tweaking.

Use Muse Glimmer. It’s good.

There is evidence that some forms of ADHD are effectively just the brain falling asleep rapidly and then waking up. This form cam be attenuated with norepinephrine reuptake inhibitor alone. Presents with ADHD like symptoms. But it’s a new thing most likely. Called “Cognitive Disengagement Syndrome”.