🔥The 9th Round of Easy Loan, Earn $40 Reward is in progress❗️ ⏰ Promotion Period: January 15th - Feburary 15th, 2025 👉 Register now and check more details at gate.io/campaigns/358

If you think reasoning is just generating intermediate tokens you don't know what reasoning is.

Its basically a breach of your fiduciary duty if you run a company and don't put Sydney Sweeney in your ads.

Don't know who needs to hear this, but materials discovery was never about randomly mixing things in a lab and happening upon transistors by chance. There are layers and layers of understanding that any ML approach could really benefit from.

pretraining is an elegant science, done by mathematicians who sit in cold rooms writing optimization theory on blackboards, engineers with total absorb of distributed systems of titanic scale posttraining is hair raising cowboy research where people drinking a lot of diet coke

tis the age of informal verification

Ants don’t build LLMs

How much of this is the result of current account deficits and recycled capital? How much due to Keynesian policies? How much locally imposed constraints on supply?

Someone should develop a self-equilibrating distributed information-processing system that takes into account heterogenous knowledge, plans, and judgment and rewards adaptation and innovation. Seems potentially useful.

Epic bacon and super grok. What a world we live in.

You worry about AGI. I worry about whatever this thing is 🤯. We are not the same.

Finance is an important functional element of the economy, directing savings to high-return use opportunities, but short-termism and myopia are problematic and sometimes industrial policy should address state-level policy goals.

Investment banking services are a commodity, full stop. There is no differentiation apart from network centrality (qua distribution reach, brand). Clients don’t pay for insight, skill, knowledge, or expertise. They pay for execution and brand halo/signaling.

Private capital markets are better than public markets, despite transparency and liquidity differences, because GPs have distinctive capabilities and returns can therefore have persistency. Private markets are probably better at price discovery than public markets too.

Capital allocation efficiency, economy-wide has profoundly benefited from the emergence and scaling of Type II error minimization intermediaries (VC) to the mix (public market investors and HFs are mainly about Type I error minimization:).

Interesting. I would want to see how this disentangles from index inclusion as a signal that the firms/sector had matured and has lower E[g] within the context of the shift from public to private capital mkts.

We may know it. But seeing it, hurts.

I mean, if we are going to think outside the box here, let’s pay suppliers in stock. Let’s allow firms to pay taxes with stock.

You see, aliefs are quick and automatic whereas beliefs are slow and deliberative. You have these two different systems: system a and system b. Human cognition arises out of these two systems. One operates ‘fast’, the other ‘slow’. One is rational, the other habitual.

Rationalist and post-rat stuff is cognitive science cosplay isn’t it?

I don’t know. This pretty much nails the business model. I’ve seen worse takes on Wall St.