Tom Reiner

Radical centrist

[Tom Reiner @TomReiner4]

1/x Twitter, Wikipedia, the internet in general, really isn't a place to get information on whole, but rather a place to get an idea of what you should research further. The spread of disinformation, misinformation, or lies, is never the true problem...

[Tom Reiner @TomReiner4]

@shaunmmaguire It is definitely a bit paranoid to think he meant, "I am going to be the dictator and no more elections" instead of, "I will do such a great job fixing the country over the next four years, that you will not have to vote to have it fixed again".

[Tom Reiner @TomReiner4]

@elonmusk "You will feel what we want you to feel, and you will like it"

[François Chollet @fchollet]

The question of whether LLMs can reason is, in many ways, the wrong question. The more interesting question is whether they are limited to memorization / interpolative retrieval, or whether they can adapt to novelty beyond what they know. (They can't, at least until you start doing active inference, or using them in a search loop, etc.)

There are two distinct things you can call "reasoning", and no benchmark aside from ARC-AGI makes any attempt to distinguish between the two.

First, there is memorizing & retrieving program templates to tackle known tasks, such as "solve ax+b=c" -- you probably memorized the "algorithm" for finding x when you were in school. LLMs *can* do this! In fact, this is *most* of what they do. However, they are notoriously bad at it, because their memorized programs are vector functions fitted to training data, that generalize via interpolation. This is a very suboptimal approach for representing any kind of discrete symbolic program. This is why LLMs on their own still struggle with digit addition, for instance -- they need to be trained on millions of examples of digit addition, but they only achieve ~70% accuracy on new numbers.

This way of doing "reasoning" is not fundamentally different from purely memorizing the answers to a set of questions (e.g. 3x+5=2, 2x+3=6, etc.) -- it's just a higher order version of the same. It's still memorization and retrieval -- applied to templates rather than pointwise answers.

The other way you can define reasoning is as the ability to *synthesize* new programs (from existing parts) in order to solve tasks you've never seen before. Like, solving ax+b=c without having ever learned to do it, while only knowing about addition, subtraction, multiplication and division. That's how you can adapt to novelty. LLMs *cannot* do this, at least not on their own. They can however be incorporated into a program search process capable of this kind of reasoning.

This second definition is by far the more valuable form of reasoning. This is the difference between the smart kids in the back of the class that aren't paying attention but ace tests by improvisation, and the studious kids that spend their time doing homework and get medium-good grades, but are actually complete idiots that can't deviate one bit from what they've memorized. Which one would you hire?

LLMs cannot do this because they are very much limited to retrieval of memorized programs. They're static program stores. However, can display some amount of adaptability, because not only are the stored programs capable of generalization via interpolation, the *program store itself* is interpolative: you can interpolate between programs, or otherwise "move around" in continuous program space. But this only yields local generalization, not any real ability to make sense of new situations.

This is why LLMs need to be trained on enormous amounts of data: the only way to make them somewhat useful is to expose them to a *dense sampling* of absolutely everything there is to know and everything there is to do. Humans don't work like this -- even the really dumb ones are still vastly more intelligent than LLMs, despite having far less knowledge.

[François Chollet @fchollet]

I like free markets, the rule of law, opportunity & meritocracy. I am not a fan of Russian-style oligarchy where the government is a mafia that uses its power to help or hurt businesses based on their allegiance (and "donations") to the boss.

[david D. @secretsqrl123]

wagner dead

russian training center hit with atacms

aircraft lost (a lot)

attacks are failing

they just bombed russia, again (russian air force)

ukr drones are tearing them up

losses are up

a senior russian stated the war is lost

POWs are up

south russia still without power

black sea fleet is only using subs

for the first time in a long time im seeing NO good news from the russian side on twitter or TG... things are changing

[Tom Reiner @TomReiner4]

@jordanbpeterson @petersonacademy and
https://t.co/TcBsVyL3JB

[Tom Reiner @TomReiner4]

@jordanbpeterson @petersonacademy Likewise
https://t.co/zuhG7RbkDR

[Tom Reiner @TomReiner4]

@ivanastradner It's already multipolar. Russia and China and several others just have not earned their place as "poles" because they lack sufficient liberal democracy and rule of law. They want fascism to exist in parity with liberty. We must now and always say, NO.

[Tom Reiner @TomReiner4]

@washingtonpost Sketchy.
https://t.co/3T7fSpGtIW

[Tom Reiner @TomReiner4]

@Rainmaker1973 "Investigators are baffled by the deaths of over a dozen people when the outcropping with a giant crack in it, inexplicitely broke away from the cliff face and crashed 604 meters down"

[Tom Reiner @TomReiner4]

@NikolaiEidissen @TheSDPadresFan @Rainmaker1973 The opposite of an unintentional jump.

[Tom Reiner @TomReiner4]

@AdrianBonenber1 I've the feeling that robot/ai inclusion is not going to go as its designers envision.
https://t.co/qoIZfIesU2

[🇪🇺Henriques 🇵🇹 🇺🇦 🇪🇺 @mendohenriques]

Campaign for @Billbrowder getting the 2024 Peace Nobel Prize. (Too late for 2023) Bill is leading a global campaign for justice since tax lawyer Sergei Magnitsky was murdered by Putin's regime of millionaires. He knows how to act: strike at their dirty money.

[Tom Reiner @TomReiner4]

@ahsokaofficial @DisneyPlus

[Tom Reiner @TomReiner4]

@HaydnJarrett @jburnmurdoch Not representative of life in what part of the UK? Or is the quality of life the same across the UK?

[Tom Reiner @TomReiner4]

@DaveQuantock @WSJ Hah, you beat me to it. If they flexed on woman's sports AT ALL, those female athletes would be selling tickets and getting those sponsorships to gain parity with male athletes. But they don't and it's all the fault of the evil patriarchy! 🙄

[Tom Reiner @TomReiner4]

@Fatherofseven7 ....this will happen and will be worse every election until the country violently entirety splits apart. Welcome to the new world.

[Tom Reiner @TomReiner4]

@Fatherofseven7 Wait until 2024. There are two opposing cultures in the US competing for dominance, each refuses to be governed by the other. Someone will resist arrest and get killed, and/or someone will lose the election, and that will kick it off and it will be much worse....

[Tom Reiner @TomReiner4]

@BrightInsight6 I see your lake Chad and raise you one trans-Saharan seaway https://t.co/JgXxmZyGrH