The multi-sensor comments are confusing. This issue is a command->semantic understanding problem, not a sensor fusion problem or trajectory planning problem per se.

It's not like the true depth of field is important for the robot to plan when it's moving at turtle speed and can stop quickly.

It's implied, and I'm hoping it's true, that this is a map-less navigation. Which is impressive. This kind of task is much easier if you have a pre-captured map of the environment, but if they are doing this without a map it's great. Historically you were always faced with "The Kidnapped Robot" problem where robots that didn't know where they were couldn't navigate even a little bit. Here the robot appears to be able to follow directions as long as they are interpretable from its current vision (or via dead reckoning).

This looks to not be an openly available model, but I think if it were, availability of an easy single-camera navigation setup could allow for a lot of cool hobbyist projects.

> achieves 76.6% on R2R-CE (Room-to-Room in Continuous Environments)

I would like to know what it did the other 23.4% of the time!

This is very cool. Congratulations to the Mistral team. Map less navigation in the outside world has been around for quite a while. But map less navigation inside the buildings is relatively new. Some stanford researchers trained a vision model (PIGEON) which could tell the geo-location from any image. It was not released publicly due to privacy nightmarish (stalking!) possibilities but I am assuming similar type of tech has gone behind this robot. if someone knows more, feel free to correct.

here's the link to the PIGEON paper - https://lukashaas.github.io/PIGEON-CVPR24/

Mistral seems to be going wide and niche. Could be a smart strategy going forward.

Ok, this is really cool. The fact that the robot can use pointing to decide where to go is a great design decision, and robotics really is the next frontier. Definitely cheering on Mistral here!

I wonder how Mistral will prioritize its robotic development against its LLM development. We have either players that prioritize both (Google, AMI), or players that prioritize coding and agentic (OpenAI, Anthropic, ...).

For a claim such as state of the art, or claims such as "great at any task" needs something of more substance. I've seen maze-solving robot competitions which can zoom around in seconds. The sped up video in the first part, and the "obstacle avoidance" are too slow for me to believe this is state of the art.

While impressive at 8B, what would the expectation be in real life, that it's run remotely or autonomously with a strapped on GPU and battery?

If you're wondering what prevents or mitigates AI hallucinations on the AI layer from replicating or acting out on the physical layer look up QNX. They manage the deterministic reasonin gof robotics. You know them better as Blackberry.

Maybe their LLMs are not the best but design is top-notch!

I love the tongue-in-cheek whiteboard mentioning Le Chaton Fat / Le Gros Chaton :)

8B sounds tiny. Of course, that's enough to easily run on device which is nice, but surely the actual SOTA must be some much bigger model?

Producing specific niche models for 100 year old industries that have mountains of data and warehouses full of folders will be the european take on AI.

It may come late but it‘ll be safe and reliable. It also requires a lot of OCR.

Robots handle clean labs well; messy real‑world environments are still the real bottleneck.

I'm ready for my home helper robot that makes dinner and does the dishes and takes out the trash.

But I'm scared for when those home helpers get drafted to fight in wars, either for or against me...

Funny how nearly all model improvements this year are demonstrated on the subset of use cases where brute force / reinforcement learning is most effective:

Robotics (using physics sims)

Cybersecurity (red team / blue team)

Math (using automated proof checkers)

Programming (using compilers)

For the record I think robotics is a totally logical place to use this training approach and this is very impressive. But if we zoom out and think about LLMs in general I’m not sure this inspires confidence in AGI arriving any time soon. I would also propose that this is a form of overfitting / training-test contamination.

Take cybersecurity for example. Through brute force techniques you will gradually memorize all of the possible exploits. So when fable breaks into a DoD network everyone is shocked but in reality it basically memorized all possible exploits including some zero day.

I’d be much more interested to see if fables performance is preserved as new exploits arise (NOT zero day - negative day meaning exploits that don’t exist yet). Would fable still find them? Or would they need to retrain it on the new software stack continuously in order to identify the zero days.

This is an important distinction that I have not seen made before.

This analysis by Toby Ord demonstrates why it’s a problem if frontier improvements are coming from reinforcement learning (brute force methods) from a purely computational perspective: https://www.tobyord.com/writing/inefficiency-of-reinforcemen...

No word on pricing or inference options i could see so not that interresting if it is not available to test.

I’m not a fan

How long until Tesla buys Mistral?

I love Uniqlo even more after seeing this.

[flagged]

Frontier labs are realizing that software/models themselves don’t have real moats and move to embodied ai.

SOTA 80% means a practically useless robot. What are they really imagining their ICP to be here?

Was it tested on a road in a car ?