Uber exited the autonomous vehicle race several years ago. AFAICT, largely because of this mismatch. Who's still in this area (in a serious way) that behaves like that?
You can request a ride via Uber. They aren't building the tech. I find it a bit unclear who is operating the fleet (Uber or Wayve), but that seems irrelevant when the question is whether the vehicles' autonomous technology is safe.
That’s a Facebook motto but Facebook (now Meta) isn’t in AVs. I suppose that excludes Facebook of pre 2016 from participating in AVs? I think everyone is fine with that.
> It was Levandowski who...had persuaded Google’s leadership to spend millions of dollars inventing self-driving cars.
> Some of the biggest fights involved risks that Levandowski was taking in self-driving experiments. ... “If it is your job to advance technology, safety cannot be your No. 1 concern,” Levandowski told me. “If it is, you’ll never do anything. It’s always safer to leave the car in the driveway. You’ll never learn from a real mistake.” [0]
It even goes so far as to call out the ones who were interfacing with AI as disadvantaged and manipulation of the advanced AI developed for the Chinese orphans led to the creation of the mouse army who were categorically trained to worship specific people as their leader which they otherwise had never known.
From what's described in the article they are using it to write the process code. That is, control code for setting the states of the SOC under test and the test equipment, then recording the data that the test equipment would measure. It can be a pain to do because each SOC and piece of test equipment might have it's own control language, drivers, different data formats, etc. This usually all has to tie back in to a local (self hosting is the norm) cloud service for quality tracking and some 3rd party enterprise software from companies like Cadence. The test equipment would be something like a Keithley Parametric wafer tester, which is actually an entire rack of equipment glued together by 3rd party code with semi-open APIs. It's not uncommon to make thousands of measurements on an SOC during verification and then process engineers work to reduce the measurement count to just the critical ones for manufacturing.
Hey, thank you so much for jumping in and for your elaboration.
> SOC and piece of test equipment might have it's own control language
OK, so apart from code generation the LLMs are providing language integration and data integration.
But how would you control quality, or make sure the LLms does not create some unwordly control code? I am trying to understand if they are using type checking or any other methodology that keeps the LLM generated code in check.
I am not from the hardware field, but I do understand that we are trying to build deterministic mashines. It is quite mindboggling that we are using statistical models for building the test suite, instead of, say, permutating through all critical input states and testing the output parameters for it.
I am asking because this isnt certainly the only mode of operation for LLMs in testing and quality control, and I would love to get a better understanding how to build such a system.
AI's can write deterministic code and have been doing so for a while. Humans do this as well. We write deterministic code with a goal, and generally it works out okay.
If you say that you can't trust AI because LLM's are statistical models, that's a true statement. Though it also turns out that we don't trust humans to write code correct either, which is why we have code reviews.
I was not trying to say that statistical can't generate deterministic code - I was wondering what tools I could use to make sure that they catch all the invariants when given a specification of inputs, logic layouts, and outputs. Certainly, feedback loops have to be involved - e.g. look at the lifetimes guarantees the Rust compiler gives - to make sure the answer is refined in iterative steps.
It's worth noting that this isn't new technology. This paper is specifically about how their new technique provides a small but statistically significant improvement on existing techniques.
The fact that they provide code and dataset is really praiseworthy.
Decades ago there was a huge push to green up cities by planting trees everywhere possible. A couple of decades after that there was a huge push to remove trees because of infrastructure and safety concerns. Now a few decade later we are back to putting the trees back in. I'm always for more trees but hopefully this time it's a little more long term this time.
This is as good as when the engineer from the Claude team said they load their website in such a way as to protect against hostile actions such as scraping.
Gotta love how they moved the "Create Email/meeting" buttons in Outlook mobile and stuck the Copilot button there so that you will hit it accidentally.
Article is short on data instead referencing a bunch of aggregated IIHS studies.
reply