@CrackedWindscreen @kentindell The #MachineLearning community needs a major reckoning in many ways (i.e. privacy, bias), but one of the *major* things that is almost never discussed is systems safety as it pertains to direct, physical harm.

For example, it is absolutely absurd that #Karpathy and others at #Tesla could basically use #IEEE #CVPR as a platform for #autonowashing disguised as legitimate ML and #MachineVision work.

No wonder the media stands no chance in grasping the real issues.

I would implore the #media to educate themselves on these subtleties for future articles.

I would also invite everyone to review Liza Dixon's seminal work on #autonowashing - a dangerous over-exaggeration of automated driving system capabilities: https://lizadixon.com/Autonowashing

Liza is on the 🐦 site here: https://twitter.com/lizadixon

Liza has always been steadfast at helping the media and the broader public understand these technologies from a #HumanFactors perspective.

That is it.

Happy #Monday!

6/6

@bgluckman Yeah... I mean... even just thinking of the "systems/industrial safety circle" on Twitter that you and I would interact with often (mostly on Tesla-oriented topics)...

Clearly, Musk is just going to de-boost all of that right away, if he has not done so already.

Try to find #autonowashing on Twitter after that.

Twitter was simply not viable for me anymore.

Why even bother was my view also.

For more information on #Autonowashing, I would highly-recommend reviewing Liza Dixon's work here (and Liza actually coined #Autonowashing): https://lizadixon.com/Autonowashing

That is it.

Appearances can be deceiving.

And the uncritical trusting of appearances when dealing with safety-critical systems is deadly.

7/7

But worse than that, the FSD Beta videos that are published, that invoke this emotional response, perpetuate #autonowashing - a dangerous exaggeration of the automated system's capabilities.

Drivers using FSD Beta become more (consciously or subconsciously) complacent and have an increasingly poor understanding of system limitations.

This, over time, will create new classes of roadway incidents - something that US roadway, in particular, do not need.

6/7