> "I get far better results in my projects by encouraging the model to argue, to push back, to poke holes in the design, to think creatively about corner cases, to be a devil's advocate, to do lateral web search to find alternatives, to challenge assumptions, to passionately advocate for what it believes is right."
> "But I don't want to engage all these assholes myself, so I spin them all up as critic subagents with another subagent to listen patiently and be the judge/arbiter."
This is the way...
No, seriously. That "sycophancy" you mention immediately after this part drove me nuts before I really understood how these things work (it's taken me a while and a lot of [painful; I hate math] research, but well worth the learning effort), but after a better understanding of the "nuts and bolts" of it all, it's fairly easy to get exactly the kinda results one should expect outta these things. If not, then "you're just holding the tool wrong". ;)
It's really just all mathematics and physics. There's no metaphysical anything about LLMs or how they do what they do. It's all just a bunch of fancy math "behind the curtain". An LLM can actually explain a lot of how it works "under the hood" if you ask it just the right questions in just the right ways. ;)
>There's no metaphysical anything about LLMs or how they do what they do. It's all just a bunch of fancy math "behind the curtain".
That's my point, but about the human brain as well. It's just a bunch of fancy math, just ones expressed with chemicals and electrical activations instead of, well, logic gates and electrical activations.
Well, I mean... Yes and no? An LLM doesn't really "think", and what mathematical fakery it does pass off as "thinking" stops the instant the text completion request finishes doing all it's math and outputting the results (as a text completion based on a simulation of a text chat most commonly). When you send it another comment or question, it starts all that math all over again, but with your new question or comment added into it's context window. It's kinda like instant amnesia each time, and behind the scenes, the software that's running the model refills it's "memory" and adds in anything new that's been added since the last prompt. But it's "memory" consists of only the "context window" it's able to handle plus the model "weights" (huge list of numbers that encode language "tokens" into a mathematical "vector space"). It never really learns anything new.
A human brain on the other hand is constantly processing 24/7 (even while you sleep), and always learning / changing until the day it dies. An LLM never changes (under the hood it's weights stay the same) unless you outright alter it's weights somehow (training / download an updated version of the model / etc). If you could somehow get an LLM to run constantly, in training mode, and give it ridiculous amounts of RAM and ultra-fast storage, and a series of fancy realtime inputs (audio, camera, etc) and maybe wheels so it could explore, and hands so it could do stuff, and access to it's own code so it could improve itself, it might eventually learn to closely approximate a really good simulation of actual thinking, but that's a bit of a scary road to go down. So many Sci-Fi movies and books end up going so very badly when the lead character starts playing in that particular sandbox. I doubt reality would go a whole lot better. ;)
> "These machines do not think and they do not have a mind."
You're so totally 1000% right about that, but they're really good at faking it, to such a degree that entirely too many people (even including some so-called "experts" in the field) have been utterly fooled by the mathematical "trickery" that performs the illusion of "intelligence".
> "The only thing you can do is to fork the conversation before it made the first mistake and give it more context or tell it to look things up."
This is a key detail that many folks don't seem to understand about LLMs in general. The generation of a response happens based on the model weights and the context window (the system prompt + everything it's fed about the conversation thus far + any additional data included as part of the overall prompt). Each response technically stands alone and is generated entirely from only that context given to it and the model's existing "token space" weights. The illusion of an ongoing conversation is maintained "behind the scenes" by keeping that "context window" updated with the current state of the conversation as context for the next prompt, but the next response is technically an entirely new generation of text.
What it all means in a TL;DR sense is that the fix for a refusal is not to continue the "argument", but simply to remove that entire interaction from the conversation entirely as if it never happened and try a different tack with new / updated / more complete context to get the response you're expecting / seeking.
I beg to differ; Most people don't deserve to be abused, but those who dish out abuse on those who never asked for it, or on other such "innocents" absolutely deserve a full measure of abuse, since they clearly don't understand (or care) how it feels to be on the receiving end of it.
This whole forum has a bunch of people who work for a guy who did the sieg heil twice at a political rally and is now a trillionaire, and the rest work for VCs and boards that have gotten rich working with him. Thats just one of them.
Remember, technology is just a tool, just like cap sheets
> "No way the US is going to nationalize a tech company regardless of what happens. The exodus of capital would be unimaginable."
You simply cannot apply any sort of actual logic to the reasoning of the current U.S. government's actions... They just "do stuff" because they feel like it, with no clear thought whatsoever of any potential consequences that may occur.
Surprisingly, I've found this works shockingly well (along with any plausible-sounding reason why it was wrong of the model to refuse) to "jailbreak" many models I've played with thus far. They're all just so eager to please...
Glad I clicked that Wikipedia link! Chinese curse... I'd always been told it was an old Bedouin curse. Learn somethin' new every day (still to this day, and every new day until I become physically incapable of learning).
> "But I don't want to engage all these assholes myself, so I spin them all up as critic subagents with another subagent to listen patiently and be the judge/arbiter."
This is the way...
No, seriously. That "sycophancy" you mention immediately after this part drove me nuts before I really understood how these things work (it's taken me a while and a lot of [painful; I hate math] research, but well worth the learning effort), but after a better understanding of the "nuts and bolts" of it all, it's fairly easy to get exactly the kinda results one should expect outta these things. If not, then "you're just holding the tool wrong". ;)
reply