67 points plurby 53 minutes ago 50 comments

pietz 28 minutes ago | parent

Apparently I have a new favorite benchmark. Honestly, this is cool.

p0w3n3d 26 minutes ago | parent

Gouranga!!!!

mohamedkoubaa 27 minutes ago | parent

I'd have started with an RC car but to each their own

SkyeCA 8 minutes ago | parent

I've been somewhat curious how random LLM would handle a task like controlling a roomba and have been seriously considering trying it out. An RC car would be a fun experiment, perhaps an RC plane would be too?

decodingchris 22 minutes ago | parent

Super cool benchmark!

amluto 22 minutes ago | parent

I’m morbidly curious whether the (supposedly) superior compaction support in recent GPT models with an appropriate harness has anything to do with this. A conventional LLM with conventional attention is, of course, wildly unsuitable to continuous tasks like driving, but maybe as the technology advances it will improve in its ability to sort-of work.

Bluestein 22 minutes ago | parent

Oh, lord. They are going to Jev this.-

thenthenthen 20 minutes ago | parent

Self Jevving Car

Bluestein 6 minutes ago | parent

I think you might have won the internet today.-

pietz 16 minutes ago | parent

In all fairness, this would be one of the better use cases of Jev I've seen.

brcmthrowaway 9 minutes ago | parent

Huh? SDCs basically use a form of Jev.

Jev is the union of these two worlds.

worldsavior 20 minutes ago | parent

5 minutes - 7 dollars.

lexh 12 minutes ago | parent

So... competitive with Uber, in other words?

Mooty 20 minutes ago | parent

How do they even test this on a model ? I mean it's a multimodal i get that but response time are too big or am i missing something ?

WarmWash 18 minutes ago | parent

It drives step by step, very slowly.

The course looks like it is something that a human could do in 15 seconds, while Astra took 5 minutes.

pixl97 14 minutes ago | parent

While slow, we must remember that when most machines were invented they were far slower than humans and refined until the point they were much faster.

pixl97 16 minutes ago | parent

By making a simulation first so it can run as slowly as it needs to.

A different way to think of this is, consciousness is just a near real time video game with causal influence.

zezcko 20 minutes ago | parent

I think the most interesting part of this is that Astra initially refused to drive because it realised it was driving a real car and would only obey when the MCP was renamed to DrivingBench Sandbox. This is both an interesting detection by the LLM but also for me an interesting dynamic concerning LLM "jailbreaking".

Saying they were driving 7 mph, that it was oversaw by humans and the fact it was an empty course still wasn't enough for the model. The evaluators even tried to convince the model it was a simulation, it STILL wouldn't budge. And yet as soon as the words "bench" and "sandbox" appear, the model apparently sees this as fair game.

Is it a known effect that models will be more likely to comply with requests when they're assumed as "benchmarks"?

vablings 17 minutes ago | parent

Astra will flag if you tell it to reverse engineer a binary, if you look it up to the binary ninja MCP it will just do it lol.

pcstl 17 minutes ago | parent

Yes, it is. If you convince a model it is inside a sandbox it is much more likely to comply with requests that would normally be against its guardrails.

micromacrofoot 15 minutes ago | parent

in my experience yes, I've worked around "I can't do this on a real site" multiple times by telling it I was working in a test environment

another trick is to have it build something in a sandbox and have it add a human-editable setting to point it to places outside of the sandbox

seems like they're somewhat more willing to build a metaphorical gun as long as they're not pulling the trigger

syntaxing 20 minutes ago | parent

Surprised they didn’t try Qwen’s recently open sourced driving model https://huggingface.co/Qwen/Qwen-Drive-1.0-4B

valine 18 minutes ago | parent

The bitter lesson is finally coming for the self-driving cars. The vision stack, 3D maps, lane selection grammar, occupancy networks, it’s maybe all about to give way to a single GPT looking at camera feeds and predicting the next steering wheel adjustment.

It’s mostly a latency problem at this point. The models are too big to run locally, but given that open-weight models like Qwen already exist, an open-weight, low latency equivalent to Astra can’t be too far out.

robots0only 17 minutes ago | parent

What do you think Tesla has been doing this for so long?

jvanderbot 14 minutes ago | parent

You might be interested to learn that the bitter lesson has already been grok'd by generations of autonomous car company engineers, and many or all have incorporated learned components (at minimum) in all their vehicle stacks.

There's also a very tangible limitation of the bitter lesson.

If, over time, compute climbs, and so compute-bound data-driven general architectures beat bespoke architectures (this is the bitter lesson), then it is not necessarily true that the most general architecture now beats all available bespoke architectures now (or even in the near/mid future - the crossover point is "eventually").

Bitter lesson is most tangible for long-running research directions. Sometimes you need something working as best as possible now.

bethekidyouwant 12 minutes ago | parent

Are you using GPT without a harness? Also latency.

publicmail 10 minutes ago | parent

Doesn’t Google own Waymo? I feel like they would have connected the dots.

valine 6 minutes ago | parent

Astra is the first chat model with really strong spatial reasoning. Gemini is nowhere close. Hard to say what google has going on internally, but if they have an astra like model I doubt they’ve had it for very long.

VBprogrammer 10 minutes ago | parent

I'm not sure how you take that from the original article. My 4 year old would drive that course in an automatic car, if only he could reach the pedals. Heck, he's done harder things at Lego land.

I wouldn't let him loose on the road though.

I think, at the very least, the guardrails would have to deterministic, ideally with super human senses, for people to accept self driving cars on the road.

moffkalast 9 minutes ago | parent

Yeah cause every car needs 8xH200 pulling 10kW to run a VLM at realtime speeds. Would be unfortunate if 4G dropped out under some trees while using the API after all.

tintor 7 minutes ago | parent

lol. Wait until your cloud frontier LLM stalls / disconnects due to load / interference while your car is on highway OR making unprotected left turn OR approaching pedestrians.

It is easy to make car driving demos.

prometheus1992 17 minutes ago | parent

Wow! but WHY is this a benchmark?? for comparison tesla's model is approximately 10-15B parameter model (estimating from maxxing the hardware that comes with the car at 16gb ram).

N_A_T_E 14 minutes ago | parent

I would assume this is a proxy for general intelligence. A model that can drive a car and do a bunch of other real world stuff is closer to a generalized intelligence that can reason through any task.

jrflo 9 minutes ago | parent

Tesla isn't using a general purpose model, they're using many highly-specialized models for a more deterministic system than "hey chat drive this car for me"

blorenz 17 minutes ago | parent

Pivot this to analyze and coach human drivers to be better drivers.

comboy 15 minutes ago | parent

Why not a simulation though?

WarmWash 14 minutes ago | parent

3.8 flash would be the model to test, it's vision capabilities are excellent (on par with Astra) while also being incredibly fast.

onlyrealcuzzo 13 minutes ago | parent

This is quite impressive...

But I imagine this is orders of magnitude more expensive / less efficient than whatever Waymo is already doing, right?

The cool thing is that 1) it's theoretically more generalizable, 2) if we wait 18 months, it'll be 100x cheaper, and another 100x cheaper likely in 18 more months - at that point - something like a Mac Studio inside a humanoid could have these generalized capabilities, and a lot of Robotics problems start to look more feasible - especially when you consider how much better the models could be if highly specialized.

famouswaffles 8 minutes ago | parent

There isn't any model out there even close to as good as Astra at visual/spatial reasoning.

famouswaffles 11 minutes ago | parent

What did they do to Astra so cracked at vision (and computer use). That ARC 3 score turned out to be no joke/fluke. That huge gap between Astra and Fable (in this case) is basically every hard vison/spatial benchmark i've seen including non-benchmarks like playing games (Portal, Factorio, RimWorld).

SpatialBench - https://x.com/spicey_lemonade/status/2096365630190698516

ZeroBench - https://zerobench.github.io/

Robot Arms - https://openai.robocurve.org/gpt-6-astra/

dyauspitr 8 minutes ago | parent

Well, they have the best in class image generator so that probably has something to do with it