mrDmrTmrJ

I used to be good a long, long time ago. (Like a decade back.) Clearly I haven't been paying attention and you're right!

Which logical steps do you most dislike? (Sincerely interested)

Regarding Meta. The author clearly states, "To wit, if you want to gain exposure to consumers, own Apple."

How does Meta's stock not give investors direct exposure to net consumer spending - via advertising and the transitive property? (I say this as someone who doesn't own Meta stock directly)

Apple and Meta are 2nd movers par excellence

Claiming that Meta's revenue is 0% consumer is something... (Why do advertisers pay Meta???)

Showing Alphabet's revenue by % of total revenue, as opposed to bar charts breaking out revenue by division, is chart crime.

I like apple and Asymco. But this is post 'How to lie with statistics 101.' Meaningfully lowers my opinion of Asymco

Do you know what a race car is? A race car is a car with, "anything and everything that wasn't necessary for that [Race] was removed to trim enough weight to make the [Race] possible."

Tesla's website literally says the Rocket version is "Not approved for public road use." And has "Seating for one." Any battery that can perform that demo can certainly power basic driving. You're accusing them of deception when every modification you describe is plainly stated on their website. (E.g. It's drive-by-wire, and the seats are modeled chassis elements. Yes, it's wildly minimal.)

The expectation that a vehicle that did a literal standing barrel roll (even if only a single time) be functionally equivalent to a Toyota Camry is madness to me. The much more interesting question is what type of localization system in 3D space were they running?

Cool - want to suggest changes to the legal code to lower the cost?

Can someone walk me through exactly how powerful distillation is?

I've never seen a good blog post or metric establishing what % of a new model's "intelligence" can be captured. I'm not disputing the document's claim. Widespread effort at distillation is clear evidence that it is highly effective. I guess I just don't have an intuitive or technical sense for how powerful distillation is.

If the leading model companies aren't profitable long-term, how do we get the money to continually spend on more research and more compute?

I get that plenty of people don't like big corporations or stock buybacks or certain CEOs, which is fine. But as someone who wants to see AGI happen FASTER, I really want to see a clear financial reason for maximal AGI investment.

If the model labs aren't clearly profitable, or open source models eat all the 'model layer' profit - what financial force will push forward very expensive experiments/scaling, etc.?

Why?

Could you write several paragraphs on this? I'd appreciate reading it.

Let's say you want to make a hybrid car lighter-weight. Where is this useful?

Power density and cycle life are truly impressive. Energy density is super low

Google's research team invented the Transformer Architecture at the center of LLM "AI". A huge invention.

California has added roughly 17GW of grid-scale batteries in the last five years. Modern GPU (NVIDIA). Modern electric car (Tesla). Reusable Rockets (SpaceX).

I agree that the US has an "Only Elon-led companies can get things done outside of computers" problem.