0c3ca83

Sure, are you volunteering your kids for the test group, accepting that if the experiment goes wrong, you can't undo it?

Yes, doing things by hand is another thing to prioritize.

A good deal of research shows that most learning happens when you recognize and back off from an incorrect solution on your own, without outside guidance. So, the key to a good learning environment is setting up scenarios where the backround information is present, and answers aren't out of reach, but there's a bit of struggle to find them.

This is also how you learn to learn, and develop grit. Grit is one of the best predictors of success.

Technology that gives answers or guides heavily is harmful to the learning process.

For responsibile learning, focus on pencil, paper, and books.

Yeah, repetitive strain injury is a common problem in the industry.

The response is also probably AI.

Am I missing something here, or is the only technical detail in the post "a trillion rows is a lot of data, maybe trying trinio will work this time"?

It feels like a lot of churn and not a lot of measurements or engineering.

The manufacturer will look at sales metrics and say "this is what consumers want more of".

They do ddos people and crawl for AI. Look at bright data and other "ethically sourced" residential proxy providers.

It comes with a botnet. You're not going to make it worse.

I'd still recommend not building one in the first place.

A computer monitor and small computer.

> just yet.

How long, in your estimates, do you think you have before your career ends?

If AI works out, possibly. The entities involved in commerce can be automated, with companies autonomously serving the needs of other companies.

I know some people who are investing based on the expectation of this being the future.

The results of the triage are the problem. Many of the triage results are invalid.

Well, we can start by pointing out that every widely available model is either directly funds or bolsters authoritarians of the worst kind.

Then we can move to the damage it does to your mental abilities.

We can continue to the ongoing effect on society, the damage to the environment, and so on.

And we can finish with extrapolating on the abilities of the models, and the dystopia that having full AGI would be. Especially if that AGI is built by the current leaders.

I think you're aware of the downsides, but you don't want to put down your precious.

They seem to believe that one. I suspect extreme harm to human well-being is far more likely.

> I'm having a hard time demonstrating the enormous activist potential these tools present

The problem isn't that they don't have potential, but that the downside isn't worth it.

You're trying to convince people that asbestos has great potential as a building material. You're saying it does an amazing job of fireproofing, and don't worry your pretty little head about cancer.

Asbestos is everywhere in older buildings because it was excellent fireproofing. We're still dealing with the consequences.

How long will it remain a thing when the voters have no leverage? King Louis didn't set up the French Republic out of the goodness of his heart. He set it up by reducing the amount of weight carried by his neck.

Voting happens because economic power implies political power. People must be kept satisfied for things to keep running.

If that deal gets upended, I don't think the economic dead weight will be treated well.

You're describing something not fit for purpose. It's not well engineered.

That's like saying "that's the irony of it, a well engineered bridge sometimes drops chunks of concrete onto cars that pass underneath".

Certainly both. About half of AI bug reports I get are invalid. Many others are questionable. It's a huge time sink figuring it out.

https://www.vulncheck.com/blog/anthropic-glasswing-receipts

You misunderstood. If machines can replace enough labor, you no longer need policy to direct wealth away from laborers.

You can just lay them off.

Yeah, so nobody did engineering.

But the experiments aren't two weeks of boilerplate. Or if they are, something else is really fucked up in the stack.

No, this is big tech slobbering at the idea of pivoting to selling AI infrastructure, and eating medium tech's lunch.

If the whole thing wasn't so harebrained, I would suspect it was a marketing campaign from Anthropic.

The irony of using AI for this...

??

But the problem identified that is talking to people and having them record data.

Like how Anthropic has started to run their own biology labs?

Today, if sanitation workers go on strike, they rapidly end up with a negotiated settlement. If we had self driving garbage trucks that didn't need them, what political leverage would they have? The same too goes for other industries.

If police have advanced surveillance technology and drones, even violent protest becomes very hard.

What pressure is there to cater to people who are powerless dead weight on the economy?

This knowledge locked in people's heads isn't going to magically show up when you throw AI at it, though.

But is it churn or value? Most of the software I use has gotten visibility worse as you deliver multiple times faster.

I would pay extra for a company that just slowed down and moved deliberately.

I always wonder why systems are designed so poorly that it takes weeks of boilerplate to use them for their intended purpose.

At what stage did the engineering fail?