News for hackers. Stuff that matters.
Re:Tesla that crashed in Autopilot mode sped up before hitti... by megakid on Friday May 25, 2018 @02:52AM (#17151236)
As an autopilot regular, I hate reading stories such as this. Just last night I had a 150 mile drive in non-perfect conditions (light rain and dusk) and had Autopilot (v1 with a single camera) engaged for the vast majority of it. It is probably my favourite feature of my Tesla. It is significantly better than Audi’s lane assist features I have used and genuinely reduces strain on long journeys. I now perform journeys without hesitation (sometimes looking forward to them coupled with a supercharger stop to recharge myself) when before I’d expect to arrive at my destination stressed and exhausted.

However, I will admit that to use Autopilot safely and effectively you MUST understand the system’s inputs, what its behaviour is based on those input (and driver controlled variables) and what its limitations are. I am a software developer by trade and whilst I couldn’t hope to replicate the autopilot software stack, I probably could write a detailed pseudo code representation of the systems behaviour (without factoring in sensor failure etc).

Re:Tesla that crashed in Autopilot mode sped up before hitti... by yellow_postit on Friday May 25, 2018 @03:36AM (#17151422)
How much do they detail those inputs, variables, and limitations vs you making assumptions about them?
Re:Tesla that crashed in Autopilot mode sped up before hitti... by ahakki on Friday May 25, 2018 @08:58AM (#17153190)
Wasn’t v1 made by mobileye (?) and teslas in house v2 has a much worse rep?