Veedrac

Wiki Contributions

Comments

This doesn't feel like a constructive way to engage with the zeitgeist here. Obviously Yudkowsky plus most people here disagree with you on this. As such, if you want to engage productively on this point, you should find a place better set up to discuss whether NNs uninformatively dead-end. Two such places are the open thread or a new post where you lay out your basic argument.

I think this is a pretty good and fair roundup, but I want to add as very lazy bit of personal context short of actually explaining my takes:

Both when I read the FOOM debate, and skimming over it again now, in my personal opinion Yudkowsky largely comes off better. Yudkowsky makes a few major mistakes that are clearly visible now, like being dismissive of dumb, scaled, connectionist architectures, but the arguments seem otherwise repairable. Contra, I do not know how to well defend Hanson's position.

I don't state this to claim a winner, and for sure there are people who read the arguments the other way, but only to suggest to the reader, if you have the time, consider taking a look and forming your own opinion.

I think signalling to someone that they've missed my intended point is most valuable if they are the kind of person to take it constructively, and if they are, I have no wish to be pointing fingers any more accusatorially than the minimum amount to bring that into focus.

I think a neutral reaction is still a plenty adequate signal in the case you mention and I value that it might do less social harm, whereas a harsher reaction is at least for me less universal as I will be disinclined to use it in prosocial interactions.

I'd be hesitant to label as "Critical" pointing out that someone has an invalid argument, and having it implicitly contrasted against "Positive" — it implies they're opposites or antithetical in some way, y'know?

Yes but the choice of word used to describe the category is not a crux. As I was imagining the interface in my head, the sections were not titled at all, and if they were titled I don't think I'd care about the choice.

Most of my comments on tone were meant to suggest better phrasings or, in the case of ‘Not what I meant’, iconography, not to suggest they were not valuable.

The specific issue with ‘Not what I meant’ is that the icon reads as ‘you missed’ and not ‘we missed’. Communication is a two-way street and the default react should be at least neutral and non-accusatory.

The section titles were meant very broadly and eg. you'd probably want to put both ‘Locally Valid’ and ‘Locally Invalid’ in that section next to each other even though the former is also Positive. To do some word association if it helps,

  • Positive — encouragement, endorsement, good vibes, congratulations
  • Critical — technical commentary, analysis, moderation, validity
  • Informational — personal actions, takes, directions, neutral

but also to be clear, I'm not wedded to this, and there are no doubt many ways to split it.

I clicked both "Disagree" and "Agree" on yours for partial agreement / mixed agreement, but that seems kind of unintuitive.

Partial agreement seems like a valuable icon to me!

Quick feedback,

  • The icons aren't all obviously interpretable to me
    • Not a crux — looks like a +, rec: ‘crux’ plus ‘🚫’
    • Please elaborate — unclear that it's a request, rec: ‘..?’
    • What's your prediction — rec: ‘
    • Examples, please — rec: ‘ex?’
    • Additional questions — rec: ‘??’
    • Obtuse — rec: remove
  • Some reactions seem tonally unpleasant:
    • Not what I meant — idea is good, but icon is accusatory
    • I already addressed this — icon is good, text could be ‘addressed elsewhere’
    • Muddled — maybe replace with ‘Unclear’?
    • Obtuse — maybe replace with ‘Too Long’?
    • Not worth getting into — feels like a meaner version of Not Planning to Respond
    • Note that I do like some critical reactions as-is, like Too Many Assumptions
  • There are too many to remember easily; perhaps
    • remove some partial redundancies, like Shrug + Seems Borderline?
    • add one-word summaries to the icon box, like [🕑 discussed 4]?
    • make it easier to see descriptions on mobile?
  • I think a top level grouping like this could make sense:
    • Positive ­— eg. Thanks, Important, Exciting, Clear
    • Critical — eg. Taboo, Harsh, Non Sequitur
    • Informational — eg. Will/Won't Reply Later, Agree to This, Shrug
  • There should be a Bikeshed emoji, for comments like this one
13742111

The only advantage of a CPU/GPU over an ASIC is that the CPU/GPU is programmable after device creation. If you know what calculation you want to perform you use an ASIC and avoid the enormous inefficiency of the CPU/GPU simulating the actual circuit you want to use

This has a kernel of truth but it is misleading. There are plenty of algorithms that don't naturally map to circuits, because a step of an algorithm in a circuit costs space, whereas a step of an algorithm in a programmable computer costs only those bits required to encode the task. The inefficiency of dynamic decode can be paid for with large enough algorithms. This is most obvious when considering large tasks on very small machines.

It is true that neither GPUs nor CPUs seem particularly pareto optimal for their broad set of tasks, versus a cleverer clean-sheet design, and it is also true that for any given task you could likely specialize a CPU or GPU design for it somewhat easily for at least marginal benefit, but I also think this is not the default way your comment would be interpreted.

A sanity check of a counterintuitive claim can be that the argument to the claim implies things that seem unjustifiable or false. It cannot be that the conclusion of the claim itself is unjustifiable or false, except inasmuch as you are willing to deny the possibility to be convinced of that claim by argument at all.

(To avoid confusion, this is not in response to the latter portion of your comment about general cognition.)

They typically are uniform, but I think this feels like not the most useful place to be arguing minutia, unless you have a cruxy point underneath I'm not spotting. “The training process for LLMs can optimize for distributional correctness at the expense of sample plausibility, and are functionally different to processes like GANs in this regard” is a clarification with empirically relevant stakes, but I don't know what the stakes are for this digression.

The mathematical counterpoint is that this again only holds for sufficiently low entropy completions, which need not be the case, and if you want to make this argument against computronium suns you run into issues earlier than a reasonably defined problem statement does.

The practical counterpoint is that from the perspective of a simulator graded by simulation success, such an improvement might be marginally selected for, because epsilon is bigger than zero, but from the perspective of the actual predictive training dynamics, a policy with a success rate that low is ruthlessly selected against, and the actual policy of selecting the per-token base rate for the hash dominates, because epsilon is smaller than 1/64.

Load More