FriendOfDeSoto

The study isn’t about how good so-called AI is at detecting cancer. The study is about how these doctors lose the ability to spot cancer after having delegated the spotting part of their difficult jobs to a model. They looked at the numbers before the introduction, while using it, and then when they took this assistance away. This study can say something about these doctors’ behavior. I don’t think it proves so-called AI is shit at it. It’s more about how humans get lazy. Roughly six generations ago people could recite poetry from memory, know the dates of historical importance, and remember 50 phone numbers. Now we’re like eff that I got those in my phone plus Wikipedia access. It’s more like that.

Not everything but a lot. The short answer is cost. This will be long and simplified simultaneously:

Ever since the latter half of the last century companies have really loved one way to reduce cost in manufacturing. And that’s labor. Go to a place where the cost of living is low and work those people to the bone for a pittance.

After WW2 a lot of stuff was made in Japan, then in South Korea and Taiwan, and then China. We have since moved on to places like Vietnam, Myanmar (when politically palatable), and India. All of these stories are different and the same. Japan’s industrial heartland was bombed to smithereens and had to be rebuilt, top of the line. People needed jobs, those people were good at it too, and manufacturing jobs went there. The economy grew, wages grew with them, and it became too costly again. Enter South Korea, after successfully democratizing in the 80s (I think). They looked at what Japan had done and did a version of that. The economy grew, wages along with it, and it also became too expensive. Enter the People’s Republic of China in the 90s, ready to blend communist political power with Manchester red capitalism. A billion people who need jobs. So they looked at what the other so-called tiger states had done and did a version of their own. The economy grew, wages grew with it, and they are teetering on the edge of being to expensive again. But their sheer size, both geographically and inhabitants-wise, keeps them in the game longer. Because the policies the communists implemented to grow and steer the economy are quite unique and perhaps the lack of having to explain everything (i.e. democratic oversight) puts them in quite a strong position. And over the last 30 years anybody who is somebody has gone to China. Big market to sell goods to, big labor force to make stuff, somebody else’s rivers to pollute. It was so tempting a deal that both the US and Europe blindly became very dependent on China. Certain base chemicals, e.g. for medicine, were almost exclusively produced there. I think there world’s entire canned mandarin industry is one village in the middle of nowhere. It takes time to change this. 47 is trying to do it the impulsive, not so thought through way (tariffs). But he may yet learn that you cannot make an iPhone in the States for the price suicidal youths put it together in Shenzhen. At the heart is always cost. Labor is expensive in Ohio, cheap in Guangdong. Slightly cheaper in greater Hanoi. If we could just stop the genocide and coups, Myanmar. India has a harder time catching up because - at least for the time being - there is democratic oversight. But the gravy train will move on. Subsaharan Africa will be the next big thing. Capitalism.

Everybody draws their own vague red lines in the sand. There is no universal law. If you like it and it doesn’t feel icky, go ahead and like it. If it feels icky, don’t. Or make sure they get no money out of your enjoyment.

However socialized your health care system is, whoever works in it is more likely than not overworked and, the lower the rank, underpaid. I feel comfortable claiming this to be true in general.

This is one study in Poland. You could draw a number of conclusions from it that isn’t just “so-called AI makes doctors dumber.” It could be just as well that so-called AI has an edge over human eyes in finding cancer. It could also be that these doctors are relieved to have a tool that’s reasonably reliable so they can focus their medicine brains on improving care in another area that this study didn’t look at. They only have so much bandwidth and they are only human too. It’s too early and too simplistic to leap to one conclusion.

And don’t get me wrong. I have no trouble imagining that the dumber conclusion could be true. Practice makes perfect so I can see if they don’t work that recognition muscle they’ll lose it. Which would eff us all when computers break or greedy companies make access to the models financially impossible. This should sound alarm bells for the educators. But once again it is too early to say the sky is falling. And it may be the pressures of the health care system that at the very least aggravate the observed effect.

Flip a coin or start both on Duolingo and see which one interests you more. This is only a hard decision in your head. If you’re not planning to move to somewhere where they speak either, this is just a hobby.

They are both romance languages so you’ll find mental handholds in either language that can help you with the other. Similar conjugations, spellings, irregularities, etc.

The French you’ll learn with internet resources or most text books will most likely be French French. As a learner, that will probably still make understanding the Quebecers an extremely hard task. It’s like somebody from a Louisiana bayou talking to a Scottish highlander. On paper, they are both able to speak English but there are accents and differences in vocabulary that increase the level of difficulty, even for native speakers.

I want to live in a world where not everything this man baby says becomes news.

I know what you mean. That’s why I see a 80/20 split there in terms of blame. Being dumb and ignorant creates mitigating circumstances but doesn’t render you not guilty by default. The flow on effect mustn’t be a complete abandonment of the concept of suffering personal consequences for doing dumb shit either. It’s a good learning case, hopefully educational for the willy nilly section of society.

If you were “shocked” that your chats were googleable, you were also revealed to be lackadaisical about your own data. Yes, they could’ve made it clearer but you yourself ticked like three boxes to get here. It’s like you accept the Ts and Cs without reading them and now the corporation owns your first born.

So why OpenAI thought this feature would be useful is astounding in itself and they chickened out faster than 47 as a result of the bad PR. But this is more on the users clicking willy nilly and not bothering to do their own due diligence. 80/20 split imo. If you thought it was okay to make your ChatGPT advice on your resume “discoverable,” you need to blame yourself.

Edited typo

I think that’s throwing out the baby with the bathwater. But regulation and hitting these companies with false advertisement penalties is something we need ASAP. And liability. If the creator of a model can be made liable for damages, that would pump the breaks on this bullshit very hard. Funny how all the so-called AI companies are averse to any of that …

Now i’m torn. On the one hand I want to dismiss your counter argument as a counter factual and therefore there is no need to even glance at it. On the other hand, the omnipotent dick is part of the equation and he can control these things and I kinda see where you’re coming from. I would say he is being more of a d though because he is rubbing Picard’s nose in how unprepared humanity is in the stars. The fact that his finger snapping detour shortened the encounter timeline with the Borgs may have had the one positive effect: forewarning for the feds. And that may have put them in a position to win by the skin of their teeth. So your counter argument holds some limited and not massive amount of water.

Q is just a dick. He could’ve brought Voyager back home but didn’t and then involved Janeway in that continuufederacy civil war. He exposed Picard and crew to the Borgs and didn’t lift a finger when he got Locutussed. I think in terms of dead people he may have screwed Picard over more. I may be wrong but I think more people died in that cross section being cut out by the Borgs as a direct result of Q finger snapping the D out to meet them. But at the core, he just does stuff because he’s a bored omnipotent dick and he makes everyone in his path suffer.

I feel like the Atari 2600 is quickly becoming for so-called AI what the “how much is a gallon of milk?” gotcha question had become for politicians who run for office. A rather pointless bit of news.

As Scotty said: the right tool for the right job. An LLM is maybe not a chess engine and that’s fine too. Why would we expect these models to be Magnus effing Carlson if they cannot reliably summarize an email or recommend eating pebbles?

It’s one study. It’s pretty sturdy in terms of methodology but it hasn’t been peer reviewed as far as I can tell. They also only looked at established software projects, not anything new. So this is a narrow scope and it doesn’t prove that so-called AI cannot enhance productivity at all. It just indicates that pro devs can be fooled into thinking they are better off with it when they are not. But I feel like that’s hardly news in these mad times.

Das war Englisch.

I think this has been true since the start. They have never not been in hot water financially.

No.

The rise of progressivism has nothing to do with corporations decorating themselves with the relevant messages where it suits them. That’s just marketing. You see that in companies who championed the marginalized during the previous administration and dropped it near instantly when 47 came in. That’s corporate opportunism.

We have seen the rise of representative democracy, of fascism, the rise of communism in the past. I don’t think we have seen anything that deserves a similar label with regard to progressivism. There is a general sine curve thru the ages of left-leaning and right-leaning politics. And thru the swings from one side to another we have still abolished slavery, enfranchised women, built social security nets, decriminalized abortion (or at least permitted it in some cases) and same sex relationships, etc. A lot of it was built on political movements but I dare say none that rose to the top and stayed there. So a rise of progressivism is as non-sensical to me as a rise of conservatism. They are just opposite ends on the political scale and we dance from one side to the other and back again.

Art is a message. It has a sender and a receiver. The sender aka the creator has an idea and their synapses create the piece of art. The receiver - even when privy to the thoughts of the creator because they talked or wrote about it etc. - consumes it and has a response. It could be along the lines the creator had intended but it doesn’t have to be. Both sides could be equally happy with their side of it while thinking completely different things.

So an artist can try to attach a certain meaning to their artwork but it is no guarantee the audience will see it that way. Is the person in Munch’s The Scream screaming themselves or holding their ears to block out screaming they hear? I read what the artist intended and I can tell you I thought the other thing.

So far I’ve been talking about a single artist and a single consumer. That’s not how this works. There could be a group who have differing ideas about the art they’re creating, like a song. So it means different things to different people on the sender side already.

It gets really messy on the receiver side because ideally the art will be consumed by hundreds and thousands of people. In that group you will have opinion leaders tastemakers and they in turn will influence other recipients. History also filters artworks. I don’t think Leo thought his postage stamp size portrait of a smirking Italian merchant’s wife would be the most famous painting in the world if experts hadn’t endorsed it, it hadn’t forcefully changed owners, hung in Napoleon’s apartment, was stolen and recovered. So there are biases built in and it isn’t as clean cut as saying everybody interprets it their own way in most circumstances.

“They say those who can’t play coach. It seems those who can’t coach stand ten rows back shirtless shouting obscenities.”

You are making it seem like this is a new problem. And it isn’t.

Centuries back it was weavers who were displaced by the industrial revolution and automated spinning machines. Coal mining went unfashionable from the late 1970s onwards and miners had to find new work. Industry in the US closed up shop and moved to China. These are just three examples of workers being made redundant in their then capacity. Two out of these three went by without much loss of life, the majority of the workforce found new jobs over time, and only some of them were screwed on a more permanent basis. Unfortunately, that’s the shitty bell curve of these changes. But another thing that’s been proven again over time is that we always think these miners or these factory workers are completely unhireable and it turns out the majority isn’t. People thought MS Excel would eradicate the entire bookkeeping profession. And they are still around and I think actually grew in numbers because they are free from pencils and calculators and could do more interesting stuff instead. Don’t fall for the so-called AI will replace everything talking point. The people who say this are either invested in so-called AI companies or drank the koolaid. All we hear for the moment is how theses models do a good a lot of the time and then break catastrophically bad somewhere. Humans still need to have a look for the time being. And thus a new job is born: chAIperone.

The problem these days is how the state responds to massive shifts like that. Social security nets have a finer mesh in the developed world outside the US. It’s much easier to go from no job to living in a car to living under a bridge in the US. A lot of people in this thread call for UBI, which is sensible but isn’t even likely in the more socialist Europe. UBI is a good answer though. Education is another one, e.g. free training programs or college classes for long term unemployed. None of that seems likely under 47.