When operating LLMs in a conversational mode, it's clear that they are deeply subject to suggestion, and spend a lot of energy just following the implications built into the questions you ask.
"Estimate this quantity." It does so, gives you a long explanation, and it seems convincing.
"Wait, don't you think this number should be higher, because of reason X?" It will say yes of course, I didn't properly consider X, here is a higher answer.
"Wait, don't you think this number should be lower, because of reason Y?" Yes of course. It will give you a lower answer.
People use these for programming.
@wookash_podcast It happens but at different levels. This hype cycle exposes the cultural flaws at organisations and flaws of individuals. The end of ZIRP did the same thing.
isnt it a bit wild that universally people in school and workplaces were taught to approach wikipedia with skepticism for potential misinformation and now everyones virtually lobotomized by gen AI chatbots that just make shit up and youre expected to use them in some cases
@cjsneo @bonzajplc I would trust them to do this in a limited capacity for specific tasks if they check in with me about it first. The problem is that these tools deeply encourage laziness and it is a VERY slippery slope til you are just telling the thing to write all the shit.
@ThePrimeagen I wonder what the sentiment will be when AI providers start charging more. Those companies are not profitable so the enshittification will come when people are addicted enough.
If you never cared about the code, you never were a software engineer.
Imagine a mathematician saying, "I never cared about math", or a basketball player saying, "I never cared about basketball, just the score."
The problem is how sensitive people have become to a good ol' tech feud. How I miss these things! I honestly think we need more of it, a lot more!
This snowflake mentality distracts us from the underlying technical debate that is so often needed. Andrew shows he still has a soul
@wookash_podcast After seeing others' reaction, it doesn't seem to be sarcastic. I'm quite surprised how people are laser focused on the ad hominem part. Could he be more professional? Maybe. But there is a lot of important information that we can take away from it.