Looks like you haven't used a decent IDE: these things have been standard for decades, locally and with minimal requirements. But wait, now it happens in the Cloud (meh, that's not gonna fly anymore, too last decade)...AND requires massive amounts of power AND cooling, PLUS it's FUBAR about 50/50.
For an incremental improvement...not great, not terrible.
I think LLMs are vastly overhyped and mostly useless, but I use Copilot as glorified autocomplete and like it.
It does what the other poster said: it automates the boring parts of "this db model has eight fields that are mostly what you expect" and it autocompletes them mostly accurately.
You're really comparing an IDE's autocomplete with something that can, at minimum, write out entire functions for you?
You're either completely misremembering what IDEs have been able to do up until 3 years ago, or completely misunderstanding what is available now. Even the very basic "autocomplete" functionality of IDEs is meaningfully better now.
Extremely lucky, probably not. More in the sense "under different circumstances, this would have gone into the folder Uninteresting Landings, never to be viewed again." Ubiquitous smartphones mean petabytes upon petabytes of boring photos and videos - the "unusual" part is just in the event itself.
So, in essence, it's now incrementally better than a templating script (except when worse), but Have Faith, it will be Better Soon. TBH, that's the same song that's been on repeat since the Dartmouth Workshop. In 1956. Jam yesterday and jam tomorrow, never any jam today.
I'd guess the law doesn't require this distinction to be reported, so it isn't :(
In my opinion - even from that definition - it still looks like level 2 autonomy: "the car sorta kinda does what it's expected, but a (remote) operator is always ready to step in and correct," whether this is "keeps the lane when driving straight" or "needs to be prodded through a left turn across traffic" is not a meaningful distinction.
A meaningful distinction would involve how much "monitoring" or "supervision" is being employed.
For example, if the robotaxi needs remote help maintaining it's lane, I'd argue it is not really "autonomous" at all --- every car would probably require 1 active, dedicated remote human "monitor". Given the inherent shortcomings of wireless communication, I'd argue Waymo crashes would be occurring on an everyday basis if this were the case.
On the other hand, if 1 person is remotely "monitoring" 100 cars waiting for a remote request from either a vehicle or it's occupant, I'd argue this is indeed a very significant level of autonomy.
The obvious problem with the original claim is that it is lacking this sort of meaningful insight.
Based on my research, Waymo claims to employ 2800 people across 6 continents to deliver 20K robo trips per day at all hours. Either everybody working at Waymo is a "monitor" or there is some significant autonomy going on to achieve this.
Level 2 autonomy is indeed sufficient for this. And where not, insurance steps in.
I mean, level 2 is great stuff, wonderful progress. Nowhere near actual autonomous vehicles though: "works well until it suddenly doesn't, and then we need someone to put it back on track."
Even a 80/20 solution operates autonomously _most of the time_, but the problem remains: it's hard to predict when it will stray from the happy path and need help.
Tesla is standing up the same type of team internally to push their cyber taxi to maintain their enterprise valuation as an autonomous vehicle company.
It turns out that self driving cannot be solved in the near term without a team of humans helping the robot drive from a far. But if you pretend you’re not a taxi company, you can command higher equity valuations.
In other words, 2025-01-01T00:00+02:00 was NOT Europe/Paris (as it was CET at that time, GMT+1), 2024-08-01T00:00+02:00 could have been Europe/Paris (CEST, GMT+2), 2030-08-01T00:00+02:00 may be Europe/Paris (CEST, GMT+2), or perhaps not (CET, GMT+1). Or it may be a completely different TZ that incidentally shares the same offset at that time.
Yesterday's "oops we bricked it remotely, no user interaction required" definitely doesn't inspire confidence - even though it was not intentional and was since fixed.
It's even simpler, I think: you're forcing it to avoid a PD negotiation, and fall back to the lowest common denominator, "500 mA on the PWR line supplied to the USB device". This goes, I believe, as far back as USB 1.1, as a slow-but-generally-safe power source for things that are barely more than "use USB PWR and GND as a dumb 5V source."
A C-to-C cable, OTOH, doesn't have this requirement, and if there's no PD negotiation, the MacBook is not required to provide power IIRC.
- note that all components need to be compliant (macbook, cable, toothbrush)
That's a lot of ifs just to charge a toothbrush. I would be greatly surprised if someone actually did (yes, it might already be cheaper to source SuperSpeed components at scale; I don't yet find it likely though)
For an incremental improvement...not great, not terrible.