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

MiMo-v2.5-Pro-UltraSpeed: 1T model with 1000 tokens per second

mimo.xiaomi.com

Post reply on HN
Print view View original

391–400 of 512 posts

Re: MiMo-v2.5-Pro-UltraSpeed: 1T model with 1000 tokens per second

#391

Post by andai »

patates wrote: post #98
cdata wrote: post #68

Earlier quoted context omitted.

We are living in a ZIRP-like era where builders at the fastest pace layer have misattributed their velocity to exponential gains in model capability. In fact, they are surfing on decades of careful effort to build a robust foundation of highly reusable software libraries. This strategy will seem to work really well until the economy that enabled that foundation to form is hollowed out. Then, there will be a reckoning…

It's not just software libraries. Specs, applications (the browser!), expectations, device integrations, operating systems, etc. So much that starting from scratch seems impossible. I'm not agreeing or disagreeing with you, but my brain cannot comprehend how machines can advance such interconnected systems while keeping humans in focus . Perhaps I shouldn't have watched the Animatrix again.

Well all we have to do is minimize animosity and ensure peaceful relations.

We're good at that, right?

Re: MiMo-v2.5-Pro-UltraSpeed: 1T model with 1000 tokens per second

#392

Post by noisy_boy »

goyozi wrote: post #58

Fast AI seems genuinely exciting and somewhat unsettling to me. Right now Claude is faster than me on some tasks but we’re at least close. I have a prompt to clean up a PR that’s been running for 1h now and I expect it to take another few. It’s hard to imagine how the workflow would look like if it was near-instant. On the one hand, it might be easier to focus. Some prompts take so long that I start to multitask and…

The first wave was just getting half decent answers. The second wave was being able to choose between actually getting reasonably ok coding results OR getting not so great results very fast. The third wave would be getting good results fast.

We need to really worry when we get amazing results very fast.

Re: MiMo-v2.5-Pro-UltraSpeed: 1T model with 1000 tokens per second

#393

Post by Klaster_1 »

8note wrote: post #327
schipperai wrote: post #161

Earlier quoted context omitted.

You can dig deeper into problems with AI. For me, it supplements my knowledge in domains I don’t fully understand. It also helps me learn. So I can tackle problems I wouldn’t otherwise. I’m excited for ultrafast AI. It likely means less temptation to multi-thread and deeper flow in single sessions.

how do you know that it is actually suggesting the right thing?

Some things are verifiable. Before coding agents, if I encountered an issue with a library or a framework, my first hunch would be to find a GitHub issue with a suggested workaround. Nowadays, I can ask an agent to really dig into it and often it does surface the root cause. For example, the other day I got a test hangup after updating to Angular 22, and the agent managed to find the bug and suggest a very trivial workaround compared to what I originally planned to go with. I reported the issue and it was fixed the next day, more or less along the lines of what I'd do.

Re: MiMo-v2.5-Pro-UltraSpeed: 1T model with 1000 tokens per second

#395

Post by addozhang »

binary0010 wrote: post #246
specproc wrote: post #112

Earlier quoted context omitted.

Yeah, flash is crazy fast, but I've found performance variable.

Flash is amazing if you know the domain really well. E.g. occasionally it makes the dumbest mistakes you've ever seen and can't correct them. However it's fairly rare, and if you know the domain really well, occasionally popping in the code and pushing it towards the correct solution takes like 20seconds or whatever. So the speed you can move with flash + high domain knowledge beats opus by a mile in my experience. I…

Yes, you don't need much domain knowledge to use Opus, but it's just way too expensive.

Re: MiMo-v2.5-Pro-UltraSpeed: 1T model with 1000 tokens per second

#396

Post by BobbyTables2 »

smith7018 wrote: post #184
throwaway67678 wrote: post #170

Earlier quoted context omitted.

Agent mania setting in It's also pretty funny sometimes how it gives weird future roadmap estimates ("part 2 - 3 weeks, part 3 - 2 months", etc.) and when you tell it to actually do those changes it's pretty much done in half an hour

I've long believed those numbers were faked by Anthropic/OpenAI to serve as a form of advertisement. The estimates are impossible to verify and their ability to do "2 days of work" in 10 minutes will presumably make the user go "Wow, I just saved SO much time!" Plus, the unnecessary text eats up the users' tokens so it helps the companies on the backend, as well.

That’s right up there with Scotty in the classic Star Trek always multiplying time estimates by 4 so he looks like a “miracle worker”

Re: MiMo-v2.5-Pro-UltraSpeed: 1T model with 1000 tokens per second

#397

Post by airstrike »

greenavocado wrote: post #380
airstrike wrote: post #360

Earlier quoted context omitted.

Funny and ironic in a way, but the point still stands that they do not actually estimate the time it will take.

> they do not actually estimate the time it will take You can't prove that )))

Right, but extraordinary claims require...

Re: MiMo-v2.5-Pro-UltraSpeed: 1T model with 1000 tokens per second

#398

Post by rstuart4133 »

throwaway894345 wrote: post #39
amunozo wrote: post #17

These price and speed optimization from Chinese providers, combined with the raising prices from American ones will change the game sooner than later. Many companies are finding issues with the AI bills already.

I wonder what are the economics driving these pricing decisions? Are the Chinese companies just subsidizing their models to a greater degree than the US, or is this an emergent property of energy policy between countries?

The Chinese economics: possibly the USA's experience.

It was pretty clear the USA won World War 2 because it out produced and out innovated everyone else. Probably with that in mind, after World War 2 the USA adopted the "Vannevar Bush" model, summarised in this picture: https://www.researchgate.net/figure/annevar-Bushs-Science-th... The idea is to jump start R&D through public funding. The hoped for outcome was that R&D feed private enterprise, leading to a productivity boom.

The boom happened, and the USA did seem to out-compete everybody else in R&D, science, and the products they delivered for decades after that.

That way of doing things seems to have faded over time in the USA. The decline seemed to coincide with the rise of Neo-econmics, and now of course it's been obliterated by Trump. He's very keen to fund Intel to produce chips in a year or two's time (which is something the stock market and banks do perfectly well), but funding basic science is getting drastic cuts.

Still other countries noticed the rise of the USA, and some adopted similar funding models for basic R&D. China seems to have picked it up with gusto, both subsidising R&D and STEM training, leading to huge numbers of engineers and scientists. Whether it will lead to an economic boom remains unknown, but acceleration of ideas and innovations coming out of China seems undeniable. More recently, Ukraine showered its local engineering garages with funds in the hopes of getting a similar outcome to the USA in WW2. It looks like it worked. If the Iran war continues, it's entirely possible arms trade will reverse: the USA could well start buying drones off Ukraine.

Re: MiMo-v2.5-Pro-UltraSpeed: 1T model with 1000 tokens per second

#399

Post by LoganDark »

I was just playing with Cerebras a few days ago because it's the fastest inference provider by far. Unfortunately, the only model anywhere near economical to run that fast is gpt-120b-oss which sucks at Pi's tool calling. So I've been hoping for something faster ever since, especially since my local hardware has a paltry 128GB of unified memory.

Hopefully this pans out and fast models (that are also not ridiculously dumb) become the norm. It's amazing what you can unlock with even a single order of magnitude's speed improvement.

Re: MiMo-v2.5-Pro-UltraSpeed: 1T model with 1000 tokens per second

#400

Post by LoganDark »

pmxi wrote: post #335
chrismustcode wrote: post #86

Earlier quoted context omitted.

You don't consider Input $0.435 Output $0.87 cache read $0.003625 per million tokens for near frontier intelligence cheap?

It’s near the frontier meaning it’s the best intelligence for the price. It’s not even close to frontier meaning it’s the best intelligence.

I hardly notice DeepSeek being inferior to Claude Opus unless I have it working on tricky and under-defined problems. That is, I trust Opus to reason much better when it has the choice. Otherwise, IME DeepSeek is far cheaper and more effective for anything where the solution is even somewhat obvious.
Post reply on HN