The fact that pirated books en-mass were used to train LLMs is legally irrelevant, oh do tell - why is that? The whole point of an LLM is to train a neural network based on content - without the content the net is entirely noise. Anthropic/OpenAI, etc. do not exist without training data. Its akin to taking millions of courses online that are intended to be paid for, but never paying.
At some point in their career the problem they're trying to solve will overwhelm even their own abilities - it just hasn't happened yet, but when it does, they'll be in for a reckoning.
All of the tooling, and processes, documentation, and code modularization that it takes us mortals to build complex things - this guy likely does in his sleep without it, until there's a problem he'll face where he can't anymore. He hasn't learned the curse of the gifted. You might have met this guy in high school - didn't need to study for his tests, aced most of his classes, then he specialized in college and ran into trouble because he never bothered to build such rigor into what he set out to achieve, it came naturally - and the longer that problem occurs the harder it becomes to learn the kinds of things that us mere mortals require.
My advice, if you can't manage him in house, manage him out
Musk bought Twitter with the same results a redneck buys a jacuzzi at the county faire - it was an expensive impulse purchase and the moment after buying it all your left with is regret
This seems very much like the episode of Veep'Boxes of lies' where they try to hide their nefarious deeds alongside the real day to day inter-workings of the vice presidents office but inundating the public with data.
Code is legacy the moment it's checked in - not writing tests is the equivalent of expecting the game of telephone to go flawlessly always and forever for every corner and major case written... This type of optimism when looked up in the OED has one definition: nope
The argument the speaker makes is extremely weak - 1.0 software programs are built of building blocks modules, 2.0 software is machine learning - what point is trying to be made here?
We still break problems down when solving problems that can apply machine learning. There's no single "drive the car" neural net, rather the task of driving a car has been broken down into subcomponents, sign detection, pedestrian detection/object in front of the car detection - and then there's logic that encapsulates these classifiers using them as inputs to determine how best to steer and power the cars drive wheels.
Its a bit far fetched to believe programming has fundamentally changed, at least not yet
If you use this to market your car without any caveats, don't be dismissal of some people who might read into that your car is capable of self-driving, when in reality its clearly absent of anything of the sort.
Tesla has likely built 80-90% of an autonomous car enough to fool those who use it into believing it can be trusted. But as with all engineering endeavors that suffer from the Pareto principal its going to take far longer to get close enough that it's better than a human whose paying attention.
Its foolish to oversell thru marketing the capabilities of AP, as it's fairly easy to conflate reliability with observed behavior.
As long as the system mimics enough of the capabilities of full autonomy and it's oversold as such customers will underestimate the risks and become victims of an unfinished product
No one wins when someone dies... I'm sure your right that Tesla employees are torn up.
My concern is that it looks like Tesla is 90% of the way there to full autonomy and the way the feature is marketing will lull even engineers who know more about how these systems work into a false sense of security and end up dying as a result -- they'll trust a system that shouldn't be trusted. There isn't a good system for detecting a lack of focus especially when it won't take more than a few milliseconds to go from good to tragic.
That's not the thrust of my point... Talking about how many people have to die to perfect autonomous vehicles is pointless, some people are willing to jump out of airplanes & they fully understand the risks.
Some number of people, N are willing to risk their lives to use autonomous vehicles they'll die as a result. It should be just as clear to person using autopilot the risks involved not misled with marketing fluff that doesn't come close to reality. Martyrs not victims
It's a deceptive to assume autopilot saves lives when it too has taken them. The number of people with access to auto pilot is far fewer to statistically determine how many more center divider deaths we might have if everyone were it's passenger.
Is the life taken by auto pilot worth less than the life taken by the aggressive driver who takes out an innocent driver? No.
I hope we eventually save lives as in net improvement in current death totals by using these technologies but the risks are not well communicated, the marketing is entirely out of sync with the risks and the "martyrs" we create thus to me look like victims.
Systems that are semi autonomous where there's some expectation of intervention work well in those scenarios, keep the car in lane markers on the highway, etc. Make sure the users hands are on the wheel but for fully autonomous even if your hands are on the wheel, how does it know your paying attention?