All ai projects should be forced to show the entirety of their training data.

Agreed—but note that in this case the information was only discovered because the organizations involved (Common Crawl and LAION) do show their data. We should assume that proprietary data sets have similar issues—but this case should be seen as an opportunity to improve one of the rare open data sets, not to penalize its openness and further entrench proprietary sources.

Yes, but in the meantime you no longer have a vehicle to get from the car dealer to the bike shop.

When

• the expected return becomes negative, or
• the risk/return ratio moves away from the efficient frontier with no other motivating factor.

The most successful applications (e.g. translation, medical image processing) aren’t marketed as “AI”. That term seems to be mostly used for more controversial applications, when companies want to distance themselves from the potential output by pretending that their software tools have independent agency.

It’s a feature, not a bug: if he claimed to know how it worked, they wouldn’t be able to sell it as a scapegoat for indefensible business decisions.

We in the West certainly need a strong defense industry.

But our defense industry sells arms to more or less anyone willing to pay. Most types of arms have basically become commodities, and the net effect of anyone producing more is that arms become cheaper and more accessible worldwide.

It’s actually easier to tell the time using the stars rather than the sun, because the elevation of the sun is hard to estimate without using a device like a sundial; but there are always stars near enough to the horizon that their elevation can be estimated with the unaided eye.

Others have mentioned geosynchronous satellites as examples that can just sit in one place, but they can only do so over the equator—not directly over North America as per your question. (They would still be visible from North America, bu they would appear to the south instead of directly overhead.) Geosynchronous orbits can be inclined relative to the earth’s axis, in which case they could pass directly over North America; but they would have to spend an equal amount of time over the southern hemisphere (without moving significantly east or west).

The Mac keyboard layout was largely inherited from the Apple ][ and pre-OSX systems, which were non-Unix. The original Unix keyboard layout differs from both Mac and Windows layouts.

The greatest insult of this month.

By next year it might be a compliment.

The assumption that, for any historical event, we have access to enough information to definitively identify the cause. This manifests in several forms:

• In criminal trials, we assume that the most likely suspect is definitely guilty. If there are nine possible suspects, and it’s shown that one is twice as likely as the others to have committed the crime, that’s considered proof of guilt—even though it’s still four times as likely to have been one of the other eight.

• In conspiracy theories, if there’s a similar situation where there are nine possible causes for a historical event, an official investigation may declare one of them the most likely—which can be true in spite of the fact that it’s still probably wrong. But conspiracy theorists, instead of accepting that that’s the best that can be done, take the improbability of the investigation’s conclusion as proof that the real cause was covered up.

• In historical study, the assumption that theories proven to be probable can be treated as proven true. A new theory might rely on two other theories, each of which is twice as likely to be true than false and is accepted as the academic consensus. But if they’re independent, the odds that at least one of them is wrong is better than even (56%).

The only likely reason anyone would bother simulating this universe would be for use as a cautionary tale, and by this point they would have already made their point and pulled the plug.

Google (and anyone else) can already do that. Reddit’s APIs provide access to the underlying database, which can include metadata not visible on the website.

I’ve seen nothing to confirm that deleting or editing your writings now will remove anything from the version Reddit shares with Google.

Break a bit off, crumble it up a bit, rinse it for a few seconds in warm water, then brew it like loose-leaf tea.

That varies by subreddit, which might actually help in training LLMs to recognize the difference.

That will remove your account from public view, but will it remove it from the data they use for AI training?

If not, you’re just enhancing the value of their proprietary data.

Sometimes there’s enough ambiguity around an event that any attempted exact reconstruction will likely be wrong in some details. Some people can’t accept that ambiguity, and take the lack of a single, definitively-proven version as evidence that the “real” version was suppressed.

We can’t necessarily judge the stability of the underlying system by the material conditions it produces at any given moment.

I thought they were fungi?