Comment Re: Everyone Is Lying to You for Money (Score 1) 38
Remington settled. They didn't admit liability.
Remington settled. They didn't admit liability.
So yeah your AI can outperform a doctor that gets 5 minutes with the patient before having to move on to the next one in order to keep their private equity Masters satisfied.
So, suppose, we stick it to the "private equity Masters", compel them to double the number of doctors — forget for a second, who is going to pay for them — and afford them a whopping 10 minutes with the patient.
ChatGPT will still beat humans... And it will be getting better with every month, whereas the humans will not...
A new study from Harvard Medical School and Beth Israel Deaconess found that an OpenAI reasoning model outperformed experienced ER doctors at diagnosing and managing patient cases
AI is sufficiently anthropomorphic to be capable of making mistakes. Demanding perfection from it is stupid. It does not need to be error-free. It just needs to be better than humans...
The Cypress Street Viaduct (the major double-decker that collapsed in the Loma Prieta quake) was built in 1957 by US contractors. Embarcadero was similarly built by US contractors in the 1960s. Russians had nothing to do with it. The only thing Russian about any of it is Embarcadero running near Russian Hill, which was named for a Russian cemetery near its peak.
because only a few at every level of government liked them *and* their legal status is very dubious
There, there. With enough of China-sponsored whipping up, the liking of a nuclear weapons research lab can be sunk overnight just as well. Indeed, this very story describes a symptom of that happening.
the rule of law is excruciatingly imperiled atm
"At the moment"? Laughing out loud...
This effectively is a fight between two branches of government, one federal, the other municipal
Federal government is at quite a disadvantage on local level — as ICE have found out dealing with other (or the same) anti-Americans.
David just might defeat Goliath
David was neither an insurrectionist, nor given aid or comfort to the enemies of his government.
I think you have the power-dynamic all backwards.
You do. Government's — even local town government's — power over businesses is immense and quite literally keeps us all from having good things.
Your data from [connected apps] isn't used to train our models
Why not — and why are people so worked up about it?
Do you resent a junior colleague learning from you too? Would you like employers to starting stupulating a right to erase memories of a departing employee upon termination of employment — lest, heaven forbid, he profits from the experience gained working at one place during the rest of his career?
There are special cases, but in general, of course conversations and collaboration should be enriching for both sides.
The thread that runs through your examples is knowingly allowing or directly facilitating known illegal activity.
It seems, you're stressing out the "knowingly" part as the distinction making a difference. But certainly, ChatGPT knew — or should have known — what the conversation was about. I've seen AI use terms like "narrative ark"...
If Google could be accused for abetting illegal drug importation, it does not seem unreasonable to go after ChatGPT in this case, not that I personally approve of either...
I asked Claude to find similar targeting of libraries or phonebook-providers in the pre-Internet era, and here are the two remotely related ones below.
The rot of criminal prosecutions of speech seems to originate from Europe...
Most of the deaths are explainable.
One suicide (Eskridge), one likely suicide (McCasland), four possible suicides (Chavez, Casias, Garcia, and Thomas), two murders (Lureiro and Grillmair), two other deaths (Hicks and Maiwald), and one missing (Reza). Neither of the murders are linked. Reza may have simply fallen while hiking and been severely injured or killed. The two other deaths were both in the age range where sudden deaths start to become unfortunately common.
It would be a problematic precedent if there were criminal liability
Here is a list of examples from the pre-AI era kindly put together for me by Claude... With the prosecuting attorneys' party-affiliations, because BeauHD felt it is important:
After claude found the issues, humans could check and fix them
Claude can do the fixing too, just as easily, in most cases. Maybe, Mythos cannot — never tried it — but Opus or even Sonnet certainly can.
Stage 3 smog alerts were year-round when I was a kid in the 1980s. They were more common in the summer, but they could happen any time the temperatures rose, and they were a fact of life at school in the spring and fall. I spent a lot of recess and PE time indoors for Stage 2 and 3 alerts. This page shows the number of days at different air qualities for Los Angeles going back to 1980. The highest number of good air quality days was 11 in 1983. For all but two of the remaining years, it was in single digits. The combined number of unhealthy, very unhealthy, and dangerous days usually covered a cumulative six months or so out of the year.
You can see the numbers shifting to the left starting in 1989. Both Republicans and Democrats in the state government (which was run by Republicans at the time) had authorized various government agencies to make changes that would affect smog levels. Since 2002, the number of moderate or good air days has covered at least half of the year, a huge reversal from the 1980s. The number of very unhealthy or dangerous air days has been in the single digits every year (bar one) since 2007, even reaching zero in 2010 and 2013 and only one in six of the other years.
It is easier to write an incorrect program than understand a correct one.