Dos anyone actually think this would reduce costs for patients by a single penny? If so, you simply haven't been watching the healthcare industry since the 1990s. The only win here is for the margins of the hospital and whatever company charges for access to the model being used. Patients won't see one penny of savings but they will see reduced quality of care as radiologists cease to exist and there are fewer and fewer actual experts to back up the AI or to ask questions of. This is just yet another announcement of another stream of money that will be going to the AI overlords instead of to actually productive humans.

Anyone who has ever commuted through urban traffic on a bicycle has been aware of just how much of an impact EVs and hybrids have been having on air quality for the better part of a couple of decades now. My asthma used to get so triggered when riding through long traffic jams on major arterial streets and I remember commenting on how that had stopped happening well more than a decade ago. And it wasn't a biological change - cycling in places like Hyderabad, India and Mexico city still did my lungs in, but I could suddenly ride through westwood and beverly hills during rush hour without any discomfort at all - because all of those idling cars stuck in traffic on santa monica blvd and Sunset aren't idling. They are powered off when they aren't moving. There's no way that can't have wider effects on air quality beyond the immediate vicinity of the tailpipes.

If they are hiring mostly early-career technologists after firing all of the long term govt career technologists, I will look forward to witnessing the inevitable age discrimination lawsuit that they will surely lose. Never mind the huge training expense of replacing experienced employees with new hires - especially new hires who will only be around for 2 years. That's about how long it takes for someone to get really good in a position/track. So why are we throwing money away by firing the people who know how to do stuff and replacing them with people who do not and who are all but guaranteed to remain for 2 years or less? This is just setting fire to taxpayer money for no reason.

Since when is 'minting billionaires' any kind of goal for anything? Why is that even remotely close to the top of a list of benefits for any change in education? Do these people really not understand that being a billionaire is utterly valueless? By definition, if being a billionaire has value, it implies that the vast majority are not billionaires. If there are lots of billionaires, then all you have is inflation. What we need are technical advancements that reduce/eliminate inequality, not to continue to search for ways in which it can be increased. The idea that there might become a class of 'education billionaires' is utterly reprehensible. In absolutely NO case should educating children be something that allows the accumulation of wealth to that degree. Nor should any other activity, incidentally. The creation of billionaires from ANY activity is, by definition, an increase in inequality that is the opposite of beneficial for every human being who isn't a billionaire. Any activity that suggests more billionaires will be created should automatically be considered deeply suspect and result in a reassessment of whether that activity has any value for the rest of us whatsoever. At some point, people are going to look around an understand this FACT. In no case has the creation of a billionaire EVER been beneficial to the rest of humanity. NONE.

You gotta love the gen-y and younger millennials who are so convinced that they've got nothing to learn from developers with 35 years of experience who wrote just about every line of code their careers are based on. I've been bumping into this more and more in the workplace in recent years. Not yet have I seen one of these arguments prove out on the side of the less experienced devs. Do you think maybe we've learned a thing or three about how to manage complex software development processes and release cycles? Like those release window rules didn't evolve out of lessons learned the hard way?

StackOverflow has been a source of terrible coding advice and overreliance on copypasta by juniors for far longer than ML has been a problem. Junior engineers have been woefully deficient in actual coding ability for (at least) the better part of a decade. No one ever reads the manual for anything, reads source code or even api interfaces, or does anything to actually understand anything about the art of actual software engineering as applied to the problems they are trying to solve. They just google search a link to an SO post and blindly copy and paste some code that is at least marginally related to the problem at hand, then rejigger things to make that solution do something useful and move on, often without testing or even attempting to understand what the code they just added even does. Now they do the same thing to ML-generated code, an outcome which I can't even tell is worse or better. At least an ML query is a firsthand answer instead of just the 4th most upvoted answer to someone else's marginally relevant question. Instead, it is a firsthand answer by an ML model that has been trained on the 4th most upvoted answer to everyone else's totally irrelevant questions.

I can't wait to see how the phenomenon evolves once enough of the code that is out there to use for training new models is itself generated by ML models. We think incest is problematic when procreating biologically...I can't wait to see what it does to incestuously generated ML offspring.

By that definition, learning is brain damage since learning alters the structure of your brain in permanent ways. Neuroplasticity is a feature not a bug.

That is an impressive collection of absolute nonsense that you wrote. Everything from the aging effects of self-medication, the impact of psilcybin and other hallucinogens, the likelihood of someone being injected with an air bubble by non-medical personnel - as if learning to inject is some kind of difficult to acquire skill, and whether addictiveness is any kind of benchmark for FDA approval - either for over the counter or prescription meds, huge numbers of which have very addictive qualities. There are over the counter allergy meds that are incredibly difficult to get off of without extreme physical discomfort (look up cetirizine addiction (zyrtec), never mind coffee, cigarettes, alcohol, cough syrup, various amphetamines, the long list of opioids used in medicine (and to terrible effect on society). Your post is just a collection of nonsense propaganda that wouldn't be out of place in a classroom during the just say no peak in the 80s.

Anyone find it not even slightly remarkable that in a world where the vast quantity of stock market transactions can now clear in a day, your bank still wants to put a 10 day hold on a cashier's check purchased from the same branch against another account at the same branch - I know because this happened to me less than a month ago when opening a new joint account by moving money from a 30 year old account. They put a 10 day hold on the deposit into the new account.

If only they had any kind of track record of criminal prosecution of greenhouse gas emitters. Sure, they issue fines and such, but no one ever gets prosecuted. This guy got hosed because he made the mistake of not being a fortune-500 company, not because he broke the law.

Companies are busted for cheating emissions tests all the time, whether vehicular or at factories and power plants, and there are almost never criminal prosecutions for doing so. Heck, just take a drive on a freeway some time and count the number of vehicles that are obviously violating emissions standards, many of which have out-of-date registration because they can't pass the inspection, so they are breaking the law when emitting excessive and illegal greenhouse gases. None of those folks are being criminally prosecuted. I guess VW executives did actually get jail sentences, but I bet those were about the fraud rather than the emissions. The coal industry has a long history of gratuitous violation of emissions and other pollution standards, largely without criminal prosecutions. I'd be very surprised if the same isn't true of pretty much every industry of reasonable size (large enough to have lobbyists, basically) in the economy. So long as penalties come only in the form of economic consequences, those consequences are just factored in as the cost of doing business and behaviours don't change other than to be more thoroughly disguised. The EPA and other regulators are thwarted from going after corporate violators both by the expense and difficulty of trying to prove individual culpability within a corporation and also by the ridiculous protections offered to corporate officers by our courts.

Congress needs to pass laws that explicitly make the c-suite employees and board members individually culpable for the actions of their companies even when they aren't explicitly aware of them or else nothing is actually going to change except for folks like the guy in this story, who lacks the protections a corporate structure would have provided him..

University educated, wealthy, upper middle class software engineers are probably not all that enthusiastic about raising their teenage daughters in A Handmaid's Tale. They likely aren't super enthused about Ted Cruz being their Senator and an economy and ecosystem where fossil fuel companies are allowed to run wild. Combine that with the cost of living in Austin starting to rival the west coast, chasing pretty much everyone else out of Austin and stripping it of its culture and there's not a lot of reason to stick around. Companies likely found that recruiting people to live in Austin was harder than anticipated once the low hanging fruit of current employees looking to escape coastal housing costs had already moved or discovered that they could as easily work remotely for anyone from anywhere in the wake of the pandemic.

Providing keys to all alameda wallets should have been a condition of his bail

It sure sounds like these carbon estimates are equating methane, generated by farmed animals that ate grasses grown in the same year they were eaten, with fossil fuel methane emitted after being sequestered from the atmosphere for tens to hundreds of millions of years. They are clearly NOT the same thing. Yes, methane is a more effective greenhouse gas than CO2, but it also persists in the atmosphere for much less time (9 years versus thousands) and, ultimately, carbon emissions that are part of the active carbon cycle aren't contributing to significantly to climate change as the crops that will grow next year to feed livestock next year will be reducing atmospheric carbon at the same rate that it is emitted, no? Cows cannot emit more carbon than they eat and the carbon they eat comes larger right out of the atmosphere. The carbon associated with fueling farm equipment and transporting livestock to market and such is relevant. Cow farts seem much less so.

why did his defense attorney need access to clearview AI in order to prove he wasn't driving? Surely all he needed was access to traffic cameras to prove that he was in the passenger seat prior to the accident? Why would someone need AI to do any kind of facial recognition to do that? Or is it that clearview has a realtime feed of traffic cameras and they needed to search for particular faces to find a corroborating image rather than just subpoenaing output from cameras they passed along their route and searching manually via mark I eyeballs? AI might have sped the process up by a few hours, but it doesn't really seem like an unapproachable problem without it.

Unless things have changed much in recent years, the default postgresql configuration is designed to run out of the box on almost any hardware - in other words, it is not even remotely tuned for real-world workloads. Granted, 1 vCPU and 2GB of RAM are pretty limited resources for a database, and performance of concurrent queries is going to be impacted by the lack of multiple cores, but there was surely much more performance to be recovered from postgres than simply increasing work_mem, never mind what would be possible by providing more reasonable resources for a db server.