naraburns
nihil supernum
No bio...
User ID: 100
naraburns
nihil supernum
16d ago
Simplicity, for starters. I've been using generative AI very lightly and I am not a coder; I still have absolutely no sense for approximating how many tokens any given task would or should actually consume.
A monthly subscription also comes with access to their "cutting edge" models; I assume you can also get that by buying tokens instead but I don't know for sure.
naraburns
nihil supernum
16d ago
My best understanding (which may be out of date) is that you'd be best served by subscribing to the $100-per-month Claude, opening it up, and asking it pretty much exactly what you just asked us. Any time you get to a step you don't understand how to perform, say, "I don't understand what this means, can you break it down more simply for me?"
In other words, I think the resource you're looking for is Claude. That's it. That's the whole list. Though probably you could do it with ChatGPT, too--I just haven't had much luck getting it to seriously drop both the sycophancy and the verbosity to focus on results. Claude seems better geared toward actual production of results.
I have tried similar things with lesser models, with worse (often, unusable) results. It will be interesting to see whether we are converging toward a single "roughly the best LLM we can create" that will eventually be baked into a chip and put in everyone's devices, and how that might impact the AI market landscape. I have a hard time shelling out even $20 per month for LLMs because I so rarely have any use for them, but in those instances where I have had a use, the most expensive models have been actually useful in ways that no other models seem to be.
I don't like the "writing science fiction" characterization because it reads as inappropriately dismissive. We're not talking about holodecks or matter assemblers; we're talking about a technology that already exists. The only speculation involved is how much better it can actually get. You could in principle train and RLHF and RAG and whatever a personally-aligned model today, you probably just don't have the resources.
Gwern is 100% correct (and far from the first person to notice) that the current business model for AI is the same as every other tech business model in the information age: to take de facto and de jure control of as much of commerce and governance as possible and then rent it back to people in perpetuity. The open source software movement was a reaction to this decades ago. The main difference at present is that AI development is a lot more difficult to do at home (though of course not impossible!), especially while the current players have a hard lock on the anticipated production runs of all the necessary hardware. Memory and storage prices have quadrupled or more; for the first time in my life, desktop computers more than 3 years old have actually appreciated in value.
I agree with the "Guardian Angel" mission as stated. It's the only real effort I've seen from anyone to align AIs properly: with individual humans, rather than with corporate or government values. But I am not optimistic about their chances.
More options
Context Copy link