The ”not perfect” part really kind of ruins it for me. I can’t trust the LLM search’s answers and have to go find the source anyway, so what’s the point?
I’m seeing people in chats post stuff like “hey I didn’t know this word also means this!” when it really doesn’t, and invariably they have just asked an LLM and believed it.
You can't blindly trust sources, either. Or, sometimes, you ability to understand the sources correctly.
I think of LLMs as bookworm friends who know a little bit about everything and are a little too overconfident about the depth of their understanding. They tend to repeat what they have heard uncritically, just like so many other people do.
If you don't expect them to be the ultimate arbitrer of truth, they can be pretty useful.
Dictionary.com isn’t likely to just outright make up word meanings. There is such a thing as a trustworthy source, even if you can’t “blindly” trust it. You can still trust it and quote it and cite it. You can’t do any of those things so far with an LLM.
The part about modern standards can't be overlooked, though. Somebody from before the 1600s would lack egalitarian values, does that failure deserve blame? That's like saying they should have invented and promoted egalitarianism early. I mean sure, in principle, but it's a lot to ask, and they should have invented transistors too.
Yes, I agree that Romans could be the least cruel and the most humane at their times actually (well, except for Christians). Still, "the Joys of discovering" feels... distasteful.
of course, losing confidence that the dollar denominated payments will be worth anything near what you expect in non-dollars is functionally very similar to losing confidence that they'll pay
Why not make something new instead? Why is the impulse to use AI to generate more of the same old stuff we’ve already seen, just with a different look?
The statements from the Finnish law enforcement agencies are basically that there is no evidence that the cables were intentionally cut. Then there’s reports citing “sources” in agencies that the incidents are thought to be accidental.
Basically, I’m not sure the reporting is properly distinguishing between “no evidence of wrongdoing” and “it was an accident”. It wouldn’t be surprising that it would be very difficult to find evidence that an anchor was intentionally dragged across a cable vs. accidentally.
Then there's China filing a patent for an underwater cable cutting technology. This can't be just a coincidence. Cable and pipeline cutting near the Baltics have increased sharply since the war started in 2022.
I am not sure, phones are a device on a network, and robocalls are an abuse of the network to get to your device. It's an intrusion made by someone else. Social media seems to think it's a network, but it's more like a bar or club with a TV in it. You choose to show up each day and watch the TV.