jacobr1

The main benefit to the fiction is limited liability. Without delegating to the corporation the responsibility for corporate actions would need to be defined as belonging to the owners, or all employees or some other structure where it all points to humans. If meta owes $18bn - I don't owe it - in the strong sense as a legal debt, as a shareholder in my 401k. The value of my ownership stake might go down (or up) but I don't have a shared legal obligation.

They still are, if what you want is roughly the same as the prior generations capability with some uplift (making then number up, but say 20% faster or more ram or whatever).

What is changing is that there genuine demand for more capabilities disproportionate to the cost decrease curve. Fab demand and supply constraints have slowed or even reversed some cost decreases - but that is still getting absorbed by the overall systems costs when you are looking at things like laptops. If you all you want is the last decades demand to browse the web and use office - things are cheaper than ever.

How much does the average knowledge worker actually use desktop apps anymore? At this point I would bet the super-majority of work is done in a browser, and secondarily electron-apps.

+fees, though ideally the fees would be capped too

Wouldn't that be covered by using PPP comparisons rather than just market currency conversion rates?

Demand for the big cities is still increasing pretty much everywhere, even if some countries are seeing net decrease in demand on the whole. Also remember that supply deteriorates. The number of "ready to move into" homes can decrease over time without maintenance and rebuilds.

Right - the point I was refuting is that LLMs are "just statistical models of words." More is going on. Does that imply "understanding?" I don't know, I'm not sure we have a good enough definition to say. But it does mean that the models are more complex that say, markov chain graphs with corpus frequencies. It seems we are encoding data in the latent space with much higher complexity than "just words." There is higher order semantic information being captured - probably not the same has human "thoughts" - but again - also not _just words_.

The surprising bit is that 10% of Americans DO NOT own a smartphone. That must be exceedingly difficult. Increasingly everyday activities require them, without fallbacks. I presume those numbers exclude children, and then if you account for the infirm, it wouldn't surprise me if we started to get to universal levels of ownership.

LLMs as clearly doing more than just modeling words. In order to predict word placement, they need to build some kind of model, their latent space, of the types of things they are able to predict. Not really full world models yet - but they have decent "blog space" or "github project" models. And you can see this with multimodality, or non-text modality modals such as images and audio. They map from their latent spaces to the outputs. The fact that multi-modal systems can share the interior layers shows some kind of internal representation is created.

The new siri models will be some variant of the gemini models. This framework seems to be more generalized than that though.

Where are the "Safe Guardrails" ?

Older ML systems were much better at exposing their internal confidence. Plenty of papers reverse out this kind of interpretability for open weight models. All the models exposed logprobs early on. This seems solvable if prioritized. The unintelligible words should be lower confidence. Getting per-token data for the output that aids with understanding the predictions is entirely feasible as engineering effort - it just won't be enough to address all the problems - but it should help quite a bit.

well, technically they can ...

I half agree. But two points: 1) if you can formalize your instructions ... then future instances can be fully automated. 2) You are still probably having the AI perform many sub-tasks. AI-skeptics regularly fall into this god-of-the-gaps trap. You aren't wrong that human-augmented AI isn't 100% AI ... but it still is AI-augmentation, and again, that sets the stage for point 1 - to enable later future full automation on long enough timecycles.