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

271–280 of 512 posts

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

#271

Post by ghshephard »

Terretta wrote: post #248
smith7018 wrote: post #184

Earlier quoted context omitted.

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.

> the estimates It doesn't estimate. It generates tokens that read like estimates associated with the context in its training material. What would you expect the generator to output instead?

I think people are continuing to view these systems as pure LLMs - when that ship sailed 6+ months ago. Between being able to review memory, using agent harnesses and sub agents and skills to go out and discover information - modern systems (Codex, Claude Code, Cursor) - use LLMs - but the LLM is only a small component of it. Compare what you get from sending a request to a chatbot like ChatGPT - to what you can from a modern harness. The output is influenced by the LLM, but it's no longer a "model making a token prediction based on training material and RLHF" - that's a very 2025 way of looking at these systems.

Even Gary Marcus is starting to come around and realize that his priors are no longer as relevant as they once were.

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

#272

Post by MaxikCZ »

holoduke wrote: post #24

Speed is indeed a next big thing what should happen with LLM frontier models. The possibilities with current models but 1000 times faster would be super useful. Earlier this week it took Claude at least full time a week with two max subscriptions to solve a complex issue where we wanted to mimic a occlusion mapping variant used in the game Crimson Desert. Pretty complex mathematical challenge. With a ultra fast LLM a…

Id also be interested in more details as sibling comment. I find that when I try to build stuff, its like building skyscraper from straw. What methods are moving you forward the most?

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

#273

Post by throwaway894345 »

orphea wrote: post #127
throwaway894345 wrote: post #39

Earlier quoted context omitted.

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?

Maybe not being led by a sociopath also helps.

I'm pretty sure Xi is also a sociopath, but he differs from Trump in that he's competent. And maybe that's a good thing for American democracy--if we had a competent dictator who could manifest massive infrastructure projects maybe the pro-democracy backlash would be significantly attenuated?

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

#274

Post by drob518 »

ttoinou wrote: post #152
dakiol wrote: post #136

So, regarding the productivity argument: I don't get it. It doesn't really matter (for regular employees) that you can do now in 2h what before it took 2 days. Why? Because it's not that you have the rest of the day for yourself. You still have to work 8h/day as usual. But now the pattern is different: instead of enjoying the craft digging deeper into problems in the span of 2 days, now you are rushing into some slot…

In which world do you live where employees work 8 hours per day ? They clock 8 hours per day maybe, but they don't work that time

I had a friend who was CEO of a startup tell me that he typically only “worked” an hour a day, not because he was lazy but just because there was so much nonsense in his schedule. He told me he was trying to get it to two hours per day.

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

#275

Post by powerapple »

dakiol wrote: post #136

So, regarding the productivity argument: I don't get it. It doesn't really matter (for regular employees) that you can do now in 2h what before it took 2 days. Why? Because it's not that you have the rest of the day for yourself. You still have to work 8h/day as usual. But now the pattern is different: instead of enjoying the craft digging deeper into problems in the span of 2 days, now you are rushing into some slot…

In my case, I think slower model makes it hard to manage context and tasks in parallel. I would much prefer to work in one task only, and finish it, take a break, and work on another task. Currently I have three tabs for three tasks in parallel, it is much worse than because constantly context switching is painful. I think a faster model would mean that you don't have to start a new task while waiting.

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

#276

Post by fullstop »

linsomniac wrote: post #217
fullstop wrote: post #141

Earlier quoted context omitted.

It's making things less fun, for me at least.

Odd, I'm having the opposite experience. The thing I really love about working with computers is when I achieve something. That's the thing that makes me figuratively, and sometimes literally, throw my fists into the air and go "Yeaaah!" With the AI tooling, I'm getting those more like a couple times a week. Plus, I'm using AI to attack the things in my day that are "a drag", and getting them done too. The highs are…

Oh, sure, I can make things with it. But I have an extraordinarily hard time saying that I made something.

It feels like it cheapens the whole thing. Maybe I'm just old, because I remember people saying the same thing about code completion in Visual Studio back in the late 90s.

This is so much more than code completion, though.

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

#277

Post by Terretta »

ghshephard wrote: post #271
Terretta wrote: post #248

Earlier quoted context omitted.

> the estimates It doesn't estimate. It generates tokens that read like estimates associated with the context in its training material. What would you expect the generator to output instead?

I think people are continuing to view these systems as pure LLMs - when that ship sailed 6+ months ago. Between being able to review memory, using agent harnesses and sub agents and skills to go out and discover information - modern systems (Codex, Claude Code, Cursor) - use LLMs - but the LLM is only a small component of it. Compare what you get from sending a request to a chatbot like ChatGPT - to what you can from…

You think someone is, or even should, special case things like estimates? What else deserves that level of intervention so they look less dumb?

Logistics for getting to the car wash next door?

In the mean time, alas, no, we can see from actual prompts sent directly or through sub-agents, and actual replies, estimates remain LLM generated.

Though, this discussion here could change that, because indeed there is a lot of special casing and context stuffing going on, one of the oldest being today's date for example.

• • •

I did read the Claude Code leak, and use pi, etc. So I disagree with your premise rather strongly. Today's "systems" remain, roughly, piles of markdown and context engineering wrapped in UI affordances, and behave very similarly today to how they did in 2024 for those already engineering context and delegating.

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

#278

Post by ai_slop_hater »

ttoinou wrote: post #152
dakiol wrote: post #136

So, regarding the productivity argument: I don't get it. It doesn't really matter (for regular employees) that you can do now in 2h what before it took 2 days. Why? Because it's not that you have the rest of the day for yourself. You still have to work 8h/day as usual. But now the pattern is different: instead of enjoying the craft digging deeper into problems in the span of 2 days, now you are rushing into some slot…

In which world do you live where employees work 8 hours per day ? They clock 8 hours per day maybe, but they don't work that time

Some companies force you to actually work 8 hours a day. It’s hell.

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

#279

Post by throwa356262 »

irthomasthomas wrote: post #11

I don't understand, given all they say, why this would not be made available to everyone at once? Why the limited release? They should have no trouble scaling it if it runs on a single rack.

The TileRT approach swaps throughput for latency, which also means less overall efficiency

Given the export restrictions this could mean they need to prioritise how to best use their limited hardware. But they could also be moving to Huawei GPUs like deepseek did and simply not have stable hardware or software for a large scale deployment yet.

This is just speculation based on the MXFP4 support on Huawei GPUs that is lacking on some nvidia GPUs.

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

#280

Post by logankeenan »

pixel_popping wrote: post #232
AlecSchueler wrote: post #226

Earlier quoted context omitted.

No, but nor can you keep track of what 10 agents are doing simultaneously. Hence the multitasking regret.

An agent can, you don't need to watch tasks, you can have a live digest with another tool.

Do you have any recommendations for a live digest tool?
Post reply on HN