3 ms·
> A recent example is people claiming that Flint, MI still has leaded water. Doesn't it? According to [1]: The residential lead service line replacement was
by nl 4y ago
> A recent example is people claiming that Flint, MI still has leaded water.
Doesn't it?
According to [1]:
The residential lead service line replacement was initially set to be finished in 2019, according to a settlement agreement with the city. That deadline was eventually pushed back to the fall of 2022 and has most recently been set for completion in August 2023, according to city officials.
and
"More than 95% of lead pipes in Flint have been replaced, and we will continue the work until the job is done," Flint Mayor Sheldon Neeley said in a recent statement on the water filters.
It sounds to me a lot like Flint, MI still has leaded water?
[1] https://abcnews.go.com/US/flint-residents-urged-filter-water-bottled-water-donations/story?id=96531880 https://abcnews.go.com/US/flint-residents-urged-filter-water...
- renewiltord 4y agoFair point. Which makes these humans (and me, I suppose) like ChatGPT: https://twitter.com/BeijingPalmer/status/1628585234865262593?s=20 https://twitter.com/BeijingPalmer/status/1628585234865262593... https://twitter.com/SCPantera/status/1127416274550546432 https://twitter.com/SCPantera/status/1127416274550546432 So ChatGPT isn't all that different from people.
- nl 4y agoThose tweets are a lot less specific though. In general the government did fix it and the water is safe. But specifically there are still 5% of pipes that have lead in them (but apparently thesr are ones that are generally not for domestic supply or something). So the exact claim matters a lot.
- renewiltord 4y agoMy comment was certainly wrong right off, though. I, a human, bullshitted just like ChatGPT.
- exodust 4y ago> "...Apple is spying on them etc. Fabrication is well and alive among us." It's more likely your claim of mass-fabrication is the fabrication. Without knowing details of what you're talking about, I'd put money on any Big Tech outfit as the guilty party when it comes to privacy mistakes, double-dipping, dark patterns, hostile nag-patterns, overreach and stealing from the cookie jar. How I wish that was paranoia and not reasoned expectation based on years of observing the behaviour of tech giants.
- renewiltord 4y agoPerhaps it is, and I'm not ChatGPT.
- Xylakant 4y agoNo, the fundamental difference is that you can reason about your mistake, can correct yourself and try to avoid falling into the same trap again. You are capable of independent thought. ChatGPT string words together.
- renewiltord 4y agoPerhaps not ChatGPT, though it does show admirable skill, but it seems quite likely that its successors may well be able to do that through sheer size. Certainly, in many conversations with ChatGPT, it is indistinguishable from a human changing their mind. Since the process of "reasoning" is not particularly well understood, it seems a bit premature to declare it different from LLMs.