angusb

I don't think there's an official divergence in terminology other than the version numbers (which have mostly stopped incrementing for FSD vehicle owners, meanwhile there is a lot of work going into new iterations for the version running on tesla's robotaxi fleet)

> Do you really want to trust...

No, but the regulator helps here - they do their own independent evaluation

> What happens when Tesla decides...

the regulator should pressure them for improvements and suspend licenses for self driving services that don't improve

agreed - I said FSD team to distinguish from "the crowds" but this was the wrong wording, should be the regulator too.

I agree it's a major mistake + should get a lot of focus from the FSD team. I'm just unsure whether that directly translates to prohibiting a robotaxi rollout (I'm open to the possibility it should though).

I guess the thing I'm trying to reconcile is that even very safe drivers make critical mistakes extremely rarely, so the threshold at which FSD is safer than even the top 10% of human drivers likely includes some nonzero level of critical mistakes. Right now Tesla has several people mining FSD for any place it makes critical mistakes and these are well publicised so I think we get an inflated sense of their commonality. This is speculation, but if true it leaves some possibility of it being significantly safer than the median driver while still allowing for videos like this to proliferate.

I do wish Tesla released all stats for interventions/near misses/crashes so we could have a better and non-speculative discussion about this!

> Your argument that a newer version is better

I actually didn't say that and am not arguing it formally - I said what I said because I think that the version difference is something that should be acknowledged when doing a test like this.

I do privately assume the new version will be better in some ways, but have no idea if this problem would be solved in it - so I agree with your last sentence.

that should have been in my list, you're right

re. extrapolation: I agree with that, but remember there's sampling error. The crashes/failures go viral but the lives saved get zero exposure or headlines. I don't think that means you can just ignore issues like this but I think it does mean it's sensible to try to augment the data point of this video with imagining the scenarios where the self driving car performs more safely than the average human driver

This has done the rounds on other platforms. A couple of important points:

- the failure here is that the car didn't stop for the bus on the other side of the road with the extended stop sign. (Obviously a kid running out from behind a car this late is fairly difficult for any human or self driving system to avoid)

- the FSD version for the robotaxi service is private and wasn't the one used for this test. The testers here only have access to the older public version, which is supposed to be used with human supervision

- the dawn project is a long-time Tesla FSD opponent that acts in bad faith - they are probably relying on false equivalence of FSD beta vs robotaxi FSD

Nevertheless this is a very important test result for FSD supervised! But I don't like that the dawn project are framing this as evidence for why FSD robotaxi (a different version) should not be allowed without noting that they have tested a different version.

BTW - I see you have a LLM answering questions based on your docs on the help pages (which is great). So really I mean for customer support issues that are raised outside this channel

Congrats, this is a great story! One small thing:

> Every time I check out competitors' sites — those who also build knowledge base or customer support platforms — I notice something odd. Almost all of them use third-party tools like Intercom or Zendesk to support their own customers. That surprises me. If your product is so great — why don’t you use it yourself? For me, that’s a golden rule: your product should be so good you want to use it yourself. If not, that means something’s wrong.

Is this not just because Intercom and Zendesk have their own ticketing systems tightly integrated to the docs? Integrating the two allows e.g. customer query auto-reply based on RAG with the documentation, or auto-replying with the 3 support articles most likely to solve the problem. I assume Perfect Wiki has no equivalent ticket integration?

Over 50% of Tesla's sales are model Y which is going through a version upgrade at the moment. The new version is widely publicised, much better than the old one, and will be available in the next few months. Why would you buy a model Y now?

A big dip was also observed during the model 3 upgrade in Q1 2024.

Everybody is obsessing so much over Elon's behaviour right now that they are glossing over important other factors.

Did anyone else read this as "Firing programmers for (AI will destroy everything)" or have I been reading too much Yudkowsky

I've been finding The Prize surprisingly light on technical details - I'm 1/3 of the way through the book and half way through the chronology of the oil industry and it hasn't even mentioned a distillation column. Does it get more technical?

I would guess that this is correct. My dreams are very much like half way between waking consciousness and a surreal movie - my consciousness doesn't seem to play a part in generating the surroundings and scenery and most of the plot, that seems to be the result of GenBrain.

I think daytime thoughts are also GenBrain btw. Only that waking brain has access to higher mental function - namely longer term memory, higher propensity to notice/surface inconsistencies etc.