imperfect_light

He's said it pretty clearly, an LLM could be part of the solution in combination with program synthesis, but an LLM alone won't achieve AGI.

You're reading a whole lot into a tweet, in his interview with Dwarkesh Patel he says, about 20 different times, that scaling LLMs (as they are currently conceived) won't lead to AGI.

The stretch was in reference to comparing Chollet to Einstein. Chollet clearly understands LLMs (and transformers and deep learning), he simply doesn't believe they are sufficient for AGI.

A bit of a stretch given that Chollet is a researcher in deep learning and transformers and his criticism is that memorization (training LLMs on lots and lots of problems) doesn't equate to AGI.

Did you listen to what Chollet said? How much of LLM improvements are due to enlarging the training sets to cover more problems and how much is due to any emergent properties?

ChatGPT agrees, "The likelihood that this was written by a chatbot is relatively high"

It should determine your adherence to traffic laws based on your social score. Musk, Thiel, Altman, Andreessen? Do whatever the hell you want, laws don't apply to you. /s

The board should have been transparent about the dishonesty, especially around financial stakes, and I suspect the OpenAI employees and MSFT would have reacted differently.

This isn't lying about product philosophy, lying about your financial stake in an OpenAI VC group would be grounds for dismissal from any company.

The denominator in this graph is houses, not people, so adult children living/not living at home would have no impact.

First of all, agreed that housing prices are insane, even two-income households struggle to buy in my city.

But taking a more long-term view, in the US the owner-occupied housing rate about 65%, which has not changed a ton in the last 50-60 years (high was 69% in 2005 and low was 63% in 1965). Granted the market has changed a lot, we have much bigger houses, and more two-income families to pay for those more expensive houses.

https://www.advisorperspectives.com/dshort/updates/2024/04/3....

We used to have housing of last resort, places like SROs that were unpleasant but cheap (and to be fair, they were often firetraps). But then we decided it was inhumane so we banned or greatly restricted them, and surprise, the people that would have been in an SRO now live in cars or tents, definitely not an improvement.

Wait, impartial is the measure now?

The most-read newspaper in Israel, Israel Hayom, is owned by a friend of Bibi. Is it impartial? Should it be shut down?

This is the libertarian argument that since businesses try to maximize profits, they would not do things like discriminate, pollute or defraud because in the long-term it would hurt profits.

The reality is businesses (including organized crime) are made up of individuals, who do illogical and short-sighted things all the time.

How is that a counterargument? "Relax, parents regularly beat their kids with sticks for centuries and they were fine.."

My first thought is that this was an attempt at satire, but now I realize it's a privileged Western person who has never traveled to a poor country and thinks that if these workers weren't labeling data they'd be self-actualized DJs or crypto investors.

So if it's mostly automated then why are they abandoning it after investing so much time and energy?

Wonder if that 30% include the cases where someone walks in, strolls around, and walks out without buying something? Which I see people do all the time, interested in the concept.

If it was just a matter of building up a large enough training set and corresponding model do you really think Amazon would be abandoning a technology they spent so much time, money and resources on, in a somewhat embarassing manner?

You can buy options (puts and calls) for three years out.

Shorting a stock for that long would be a recipe for disaster, at some point it's going to go up and you're going to lose it all. Remember when you buy long your positive upside is infinite and your downside limited to 100%, when you short it's the opposite.

>So the market price for a company tends to favor short-term cash extraction over long-term value creation.

Then why did markets favor companies that literally made no profits for years, Amazon and Uber being examples?

More broadly, what percentage of mergers are successful? I'd guess less than 20% if compared to the analyses that make them attractive to companies.

I've been involved (at the worker bee level) in multiple mergers and in every case where they attempted to merge systems it easily took 5-10X the estimated time, cost and resources.

The most successful mergers were the ones where they either left the companies separate (minimizing the ability to reduce costs) or the ones where they just threw away one of the company's systems, got rid of most of their employees, most of their products, and just acquired their customers.

The variation in price is crazy, $80/MWH for Atlantic vs. $20-$30 for many other regions?

It's because in many of these fields (medical, dentistry etc) professionals are making decisions based on intuition and what they learned in school (which may or may not considered accurate anymore).

They should be consulting a formal decision tree based on latest research, but many doctors consider that to be insulting, much like they pushed back on check lists when they first became a thing.

I wish that "working remote" translated into "live and travel wherever I want" for everyone.

So many companies now have policies that restrict where you work from for tax/payroll purposes, or because they don't want employees working in country where the employee doesn't have the proper visa (even if working for a US company).

The problem is that it's become so difficult to tie into the grid that projects are being cancelled.

There was a NY Times article on this (http://web.archive.org/web/20230226032242/https://www.nytime...) that mentioned that the PJM Interconnection, the biggest US grid, is not even accepting new applications for large projects until 2026.

This never made any sense to me, aren't you just guaranteeing that the most marketable employees will leave and those that are coasting will stay put?

This is like the old "athletes are too greedy!" as if the alternative is the team owners (well known for being non-greedy?) should keep that money instead.

The title might lead some to believe these planned buildings have affordable housing, I don't believe that's accurate. Instead they will pay the city for the right not to have any affordable units, hence the $39 million.

Seattle notes:

Escala condos were the model for the protagonist's apartment in 50 Shades of Grey

https://www.mansionglobal.com/articles/fifty-shades-seattle-...

Also across the street from Lola's, where Bezos often has breakfast.