I suspect it's going dormant for a couple years and then he'll say "Hey, this robot thing isn't working out, so we're closing the facility." He doesn't have any desire to stay in California.
I had the same few years ago. When I pointed out that I can get full root with most of the whitelisted commands they answered "We know. It's not about security but to prevent lusers from accidentally rm -rf /* the server. Feel free to spawn a root shell. You obviously know what you do"
Biometrics were a very crowded market during the 1980s and 1990s when it was a newer idea and electronics were starting to make things practical. Lots of ideas were tossed around before the industry pretty well consolidated on fingerprints with a side of iris imaging and hand geometry in some more security-sensitive niches. It mostly came down to cost: fingerprint scanners, even before the modern capacitative type, came down in price much faster than other types of imaging (visible rather than IR sensors, glass platen allowed for fixed focus, etc). The widespread use of fingerprint comparison in criminal forensics also mean that there's an older and stronger academic literature on fingerprint comparison, whereas other types of biometric sensors often involve proprietary match algorithms and you have to rely on the vendor's assertions about reliability.
Of course everything around cameras has come down in cost tremendously since then, so palm imaging is probably reasonably priced now, but it lacks a clear enough advantage over better-established methods for anyone to switch over. Besides, just the fact that you have to position your palm the way you do makes it difficult to install them in most practical door situations. Fingerprint sensors turn out to be very compact and fairly intuitive to use.
I scoured Amazon's sales materials around Amazon One very closely, because I found it fascinating that they were seemingly trying to revive the technique. I was surprised they were doing it as a payment device, but it made more sense when I found materials (I think old FCC filings) that suggested that it was originally designed as an access control product and perhaps "pivoted" to payments later. The strangest thing about it though was how unconvincing the sales materials were, it felt like they were really grasping at straws for a reason to select it over other options.
From what I could find it doesn't appear to have been an acquisition; the regulatory paperwork was all filed under some LLC but it seemed to just be a front company for Amazon which is fairly common for that kind of thing. So my best guess is that it was a pet project of someone influential enough to burn some R&D on it, and maybe pivoting to payments and putting them in Whole Foods was thought to maybe be the hail Mary that would turn it into a real business.
The actual integration with the PoS in the stores was clumsy too, they Velcro'd and NFC antenna to the side of the credit card terminal to use to make payments by proxy card. I originally got obsessed with it because I was trying to ID the suspicious device Velcro'd to the payment terminals at Whole Foods!
This is very compellingly written but I've found a lot of it not to be true for me. It is gray and rainy and a bummer outside most of the time, so frequent sights of that are like frequent sights of politics news. In my home office, I don't have much of a (detectable) background thread running for who could be at my office door, but opening the window and seeing people walking their dogs etc directly at the edge of my unfenced front yard absolutely causes that sense to flare up, even if they're all harmless and I recognize half of them.
Getting the room cozy and psychologically satisfying was a huge deal, and I"m really glad I did it, but the end result is much closer to horse blinders. I have ADHD, so distraction minimization is the name of the game.
I agree with this. Hopefully they're able to track down who did this. To upload to ADS-B Exchange you need an account. But it's not that difficult to get one. I'm not sure what kind of information they may be able to get on it. As you say the person who uploaded this may not be anywhere near there. The aggregators probably should have heuristics like if only one feeder in an area with a decent density of feeder coverage uploads an anomalous track, it should get flagged.
I have a nasty suspicion that far fewer of them will be, that CS and SE based professions will end up collapsing and consolidating into a handful of AI megacorporations and a guild-like elite of AI-herders will be what's left.
Cool stuff. I just started open sourcing my code for running Django apps in production [1]. It's a command-line tool that handles deploying to a server, SSL certs, databases and backups, automatic error emails, and background tasks via celery / redis. The best part? It does not need Docker. It just runs everything on bare metal.
I'm guessing by selling fine-tuning, consulting on hosting, and other services? They also seem to be offering their own inference service with their model, obviously as an open weight model that will be commoditized but I'm sure there are some people who'd prefer to buy from the originating lab. But yeah, when you're offering open weights models, your customers are going to be people who want to self-host, fine tune, etc, so they might be offering services for that.
I mean... you can't think of any ways that AI could actually generate new value? Or more abstractly, of a way that Jevons' paradox can't apply in the case of AI?
