Can't I just parallelize my LLM calls myself for each question?

You can. It will be expensive, slow, and less reliable than a specialized model.

Anything you can do in Jev can be done with an LLM at much greater cost and latency.

Agree, except the probabilities for outcomes in the structured output. I don't think you can get those for most frontier LLMs (logprobas). You can get it for open source models but not frontier LLMs.

That number is a big deal, assuming it is well calibrated. Did they talk about calibration?

I've seen them talk about it a bit on Twitter -- it seems to be fairly well-calibrated in general, but obviously you need to test it on your use case and dial it in comparison with known data for best results.

Maybe you need Prolog instead.

Controlling critical technology your software runs on.

Indeed. People literally make stuff up when their corpus callosum is severed.

Aand what if their corpus callosum isn't severed?

They also make stuff up.

> People have an inner monologue

Perhaps up to 50% of people actually don't have an inner monologue, much like many have aphantasia where they can't actually see anything in their mind either.

That’s not how that works. I don’t usually have an inner monologue either, but I do have an abstract stream of thought. It’s not as if I am always acting on instinct.

Sometimes this is useful so I know it didn't accidentally commit something I didn't want it to.

The site is opusfived.com, and Opus 5 is probably the worst so far at doing this.

I dunno, in my experience it's often overly-narrow, sometimes jumping through all kinds of hoops to preserve some edge-case behaviour that doesn't matter because I didn't mention it could be changed.

Experiences vary, yes.

But you can see in this thread that folks definitely have experienced this.

This is exactly the experience I have with Opus 5. Opus 4.6 is better, Flable 5.1 much better. But Opus 5 is infuriating.

Just download the ChatGPT app and use the Codex mode.

Pascal's AI, I see. But yes, social media algorithms and the Internet in general have made people realize doomerism is what gets the clicks and views.

I mean, kind of yeah? But like, there are already people working on the chips, there are already people designing the next hardware, etc. I'm sure we're going to get to a plateau in capability soon-ish? Exponentials are actually all sigmoids. But what does that look like? If the plateau in capability is 100s of times smarter than the average person (arguably we're already there in many many but not all domains) then in 10 years time last year's reasoning model etched into an ASIC or some crazy monstrosity built out of FPGAs but for LLMs is probably way more than enough for 99.999% of use cases?

But yeah, doomerism is the dominant narrative of the day here right now. There's a sort of eschatological poisoning that's happening presently. Nobody can even seem to imagine a world where things get better. It's crazy. Maybe it's because I recently went through a major illness, maybe it's because I hit my head one-to-many times along the way? But I've never been more optimistic about the future than I am now.

You really think Hume was the first one to think of this in hundreds of thousands of years of human history? As a very easily findable counter example, you have Al-Ghazali who came up with this several hundred years prior.