>"Earlier this year, when researchers asked 10 leading chatbots about topics targeted by false Russian messaging, such as the claim that the United States was making bioweapons in Ukraine, a third of the responses repeated those lies."
I spent my high school years on 14.4, 28.8, and finally 56k modems. When I got to college, I got my first taste of broadband, and boy did it get me in trouble. I would be up all night trading files on mIRC.
It already is for me. I've been using LLMs daily for years now. I don't get the people claiming AGI every two minutes any more than the people claiming these tools are useless.
LLM reasoning abilities are very fragile and often overfitted to training data. But if still you haven't figured out how to do anything useful with an LLM, warts and all, that says more about you than LLMs.
I don't believe LLMs will directly lead to AGI. I'm also annoyed by the folks who hype it with the same passion as crypto bros.
As new "thinking" techniques and agentic behavior takes off, I think LLMs will continue to incrementally improve and the real trick is finding ways to make them work with the known limitations they have. And they can do quite a bit
It’s the flip side of taking kindness for weakness, and in my experience people who do that reveal something concerning about their own character and worldview.
> Now, electricity is not the boogeyman: you can open a power supply safely, you just have to be aware of the dangers inside. First, never open a running power supply: if you accidentally touch something at high voltage, you could die.
Fortunately it wasn’t fatal, but I stupidly did that as a kid. It was my first PC build (an Athlon Thunderbird system) and it was crashing. I fixed the problem by adjusting the PSU voltage rails with a screwdriver while the system was running so I could see the changes in the BIOS.
- Eva Cassidy, “What a Wonderful World”
- Leonard Cohen, “You Want It Darker”
- Camille Saint-Saens, “Danse Macabre”
- Paul Simon, “Diamonds on the Soles of Her Shoes”
- DubXanne, “Roxanne, Police in Dub”
It’s this one. And it happens in every industry. Things fall through the cracks because no one wants to take responsibility for covering the cracks. And that’s because it’s generally a thankless job. The company Swiss Army knife / fixer / keeper of institutional knowledge is quietly holding things together while the upwardly mobile types ensure the spotlight stays on them.
Moving fast also shields you from accountability but allowing you to accrue the perceived gains, even if they don't end up existing (again, lack of accountability)
Wood frogs can survive freezing solid. Their liver produces glucose to flood all cells, prevent cell freezing, and protect against dehydration. Ice forms around cells and organs but not inside them, preventing lethal damage.
And if the book was about how such a "virus" might work, the author could have brought that up. But since the book is actually exploring how different cultures or people groups might react to the same event, going down that road would be a trap, and only invite more intense scrutiny from pedants. Explicitly "lampshading", as a sibling commenter put it, is the better way to achieve the author's goals.
Broader than that, it’s critical thinking skills. Using search and LLMs requires analyzing the results and being able to separate what is accurate and useful from what isn’t.
From my experience this is less an application of critical skills and more a domain knowledge check. If you know enough about the subject to have accumulated heuristics for correctness and intuition for "lgtm" in the specific context, then it's not very difficult or intellectually demanding to apply them.
If you don't have that experience in this domain, you will spend approximately as much effort validating output as you would have creating it yourself, but the process is less demanding of your critical skills.
No, it is critical thinking skills, because the LLMs can teach you the domain, but you have to then understand what they are saying enough to tell if they are bsing you.
> you don't have that experience in this domain, you will spend approximately as much effort validating output as you would have creating it yourself,
Not true.
LLMs are amazing tutors. You have to use outside information, they test you, you test them, but they aren't pathologically wrong in the way that they are trying to do a Gaussian magic smoke psyop against you.
Knowledge certainly helps, but I’m talking about something more fundamental: your bullshit detector.
Even when you lack subject matter expertise about something, there are certain universal red flags that skeptics key in on. One of the biggest ones is: “There’s no such thing as a free lunch” and its corollary: “If it sounds too good to be true, it probably is.”
>"Earlier this year, when researchers asked 10 leading chatbots about topics targeted by false Russian messaging, such as the claim that the United States was making bioweapons in Ukraine, a third of the responses repeated those lies."
reply