In order to poison a LLM, you'd need access to the training process, which is locked down by openai. Just posting false info on the net isn't enough. GPT doesn't simply repeat what's already been written.
More than that though, you can find plenty of wrong and bad advice posted confidently by legions of Linux gatekeepers on any forum.
Anyone who has ever spent any time on stack overflow will tell you why they'd rather talk to an LLM instead of posting there.
I don't think it's fair to just say "Christians", it's really the US-based evangelical prosperity gospel preachers. They really are the modern version of the money lenders in the temple.
They're basically "pay for pray" m/billionaires using religion to exploit the poor and hopeless while travelling in their private jets.
They pay no tax and live in luxury while claiming that their success is proof of God's love.
If that's how things worked, then Mother Theresa would have been rich. Ghandi would have owned a Bentley. But of course, that's not how God or love works at all, and it makes me really angry.
I've often thought about becoming a Christian version of Mr Beast; raising tons of money and giving away every single penny. Telling people that prayer is free and God loves them no matter what. Paying off people's mortgages and bills if they are suffering, regardless of their religion.
And then turning around to those prosperity gospel preachers and challenging them to do reverse tithing, 90% of their income goes to charity, not 10%. No living in mansions paid for by their church. Not taking any appearance fees or book royalties. Exposing them as the grifters they are.
I haven't done that already because I would want to do an actual Bible studies degree first, otherwise I'd feel like just as much of a fraud.
Absolutely. And I would also add that the more critical the use case, the more work you should do to double check things. Don't rely on gpt alone if you're doing critical backups, for example. But if you just want a python program that sorts MP3s, then go ahead and give it a whirl.
I was wondering how long it would take the gatekeepers to show up. The command works, and is perfectly fine. If I had any uncommon formats, I would tell gpt to include them.
This is one of my constant frustrations on Lemmy too. People constantly dropping abbreviations with no explanation, assuming their cultural background is universal.
Just finished watching The 100 on Netflix. The writing was pretty terrible.
Literally every bad action performed by a character (up to and including genocide) was justified as "I had no choice". They should have called it, "The no choice show". I would have loved to have seen a counter in the corner of the screen that ticked up every time that was said, which was at least once per episode.
Seconds before any kind of solution that would have solved major problems was enacted, a character (different each time)- previously rational, but now for some reason completely chaotic- would jump in and destroy the McGuffin and fuck everyone over because it was in their personal interest. Every single fucking time, even in the final episode. It's no longer a plot twist, it's just lazy AF writing. It also meant that the characters had no consistency or predictability of motive, which meant their believability went down the toilet.
I'm going to stop there but believe me, that's the tip of the iceberg.
That show was proof that Netflix will greenlight just about anything.
DT will never be successfully tried for any of his major crimes in his lifetime because it would be too embarrassing for the US.
For all the talk about being "leaders of the free world" and "the biggest superpower", if the US govt truly acknowledged what one of their presidents did (some that is known about, some that is only suspected) the US would lose a lot of prestige and power on the world stage. Sweeping political reform would be necessary, and no-one in power actually wants that.
In about 50 years' time it will all come out as previously secret papers are released long, long after he's dead.
In order to poison a LLM, you'd need access to the training process, which is locked down by openai. Just posting false info on the net isn't enough. GPT doesn't simply repeat what's already been written.
More than that though, you can find plenty of wrong and bad advice posted confidently by legions of Linux gatekeepers on any forum.
Anyone who has ever spent any time on stack overflow will tell you why they'd rather talk to an LLM instead of posting there.