Decompiled

[Decompiled @decompiled_dev]

@callebtc If it weren't for the patent on it, schnorr sigs could have be used much earlier and for a lot more things.

They were discovered many years ago, but implementation was slow due to legal reasons.

[Decompiled @decompiled_dev]

@TheBlackHorse65 Using moral language for personal benefit is one of the defining traits of "Suicidal Empathy". John 12:6 is a clean example for why people engage in it.

Scripture states the incentive, but I see both describing the same political malady, moral language used for self-interest.

[Decompiled @decompiled_dev]

@callebtc Cryptography is used in some shape or form on every website. One of the most widely used technology which shapes the entire internet. I am very grateful we have many open source implementation to use.

[calle @callebtc]

Cryptography is so incredibly fun.

As a humble physicist, I never had to study discrete math until I started working on Bitcoin. In the beginning, I didn't even know what an elliptic curve was. I had no idea what a public key is other than its purpose in encryption.

Nowadays, I think cryptography is the most fun application of math that I've ever worked with. It's truly insane what you can do with numbers.

Hiding information, agreeing on rules, committing to actions, forcing outcomes. So many crazy things are possible just by putting a bunch of numbers in a jar and mixing them really hard.

Study this shit if you can. Life is short. This is one of those human technologies that will blow your mind once you grok it.

[Decompiled @decompiled_dev]

@TheBlackHorse65 UK PMs are like ablative shielding. A new one will be installed when the old one gets worn out.

[Decompiled @decompiled_dev]

@asymmetricinfo Journals have too many pay walls to be read outside academia.

[Decompiled @decompiled_dev]

@Oblivious9021 Only relevant context helps. If you add random info there is more for the attention to find what's relevant.

More context = more cost.

I am of the view using the least amount of context for the task leads to better answers, but its nice to have more available when you need it.

[Decompiled @decompiled_dev]

@ToKTeacher I agree. I thought this was basic science.

Every experiment needs a falsifiable hypothesis before you start. Experiments rule out competing explanations, but by themselves don't provide understanding.

[Decompiled @decompiled_dev]

@DaveShapi I like the part where he says using soft laws through agency FUD is expected.

Make up a reason to scare enterprise into paying, but not too scary to avoid push back.

I think his problem is dismissing other views so he can be the smartest guy in the room.

[kepano @kepano]

formats evolve over decades, apps over months, intelligence is evolving every day

in terms of pace layers "file over app" defines priorities: invest deeply in formats, and don't expect anything at the intelligence layer to last very long

[Decompiled @decompiled_dev]

@stevewdavens @EMostaque @grok The open weights discourage funding for bigger and bigger training runs. If you are hyper focused on frontier labs the decel arg makes sense.

Personally I think it means more companies are going to invest in their own hardware, which out weighs the loss of moat for the big labs.

[Decompiled @decompiled_dev]

@SyarmineS @anwaribrahim He has a book where he lays it out. 0 loyalty to the host country is a feature.

Tons of misrepresentation of Western philosophers too. He just takes their names, misquotes and misrepresents them to lend faux authority to whatever point he's making.

[rohit @krishnanrohit]

Here's the steelman argument to be anti open source AI as it stands today (note: I don't believe it).

The things you need to believe are that AI requires large scale capex as a necessity, and the open source movement is basically all akin to piracy, just duplication without effort.

So the logic has to be:
1. Frontier AI requires large Capex, which depend on their ability to capture huge economic rents
2. OS AI today mostly is equivalent to piracy of what frontier AI does
3. It contributes little added capability or research, it will not expand the market or increase demand
4. It materially decreases the rents that would finance point 1
5. Gains from OS, access, competition, innovation, deployment, downstream application and benefits, are much smaller than the costs of lost frontier lab R&D

And *if* this is the case, that larger AI labs are like pharma cos, then sure we should have the governments step in and license every part of that supply chain, with extreme restrictions on usage ("behind the counter", "prescription drugs").

Now, the reason this analogy doesn't work in reality is that OS AI models are not just duplicates, generic knockoffs of Claude. If I create knockoff Xanax, Pfizer loses and might invest less in new drugs, but the knockoff Xanax makers aren't making better alprazolam, much less other drugs. They're not improving the molecule, but OS model makers are!

Giving folks an open model isn't like handing them a knockoff iPhone to use, it's capability to search the entire adjacent design space, a distributed search process.

This gives tremendous option value, since it is not clear where the next best ideas come from. Maybe it's OAI spending $10 Trillion on a training run, but maybe not!

The purpose of the acceleration-is-good thesis is not that giving OAI and Ant a de-facto monopoly will result in larger models, which will bring about societal largesse, but that this technology is useful and good, which is why more people building and developing it is useful.

I wrote this because the quality of discourse on this is terrible, most advocates for both sides don't seem to be able to actually explain their worldview. If the best argument for why to ban/control OS is Ant needs an extra trillion for a larger pretraining run that's yet another marginal improvement on the same capabilities line, then the market ought to decide if that's worth it. If it sees enough value, it'll fund it. If it doesn't, it won't. And that's fine.

[Decompiled @decompiled_dev]

@vesperdigital @TheAegeanMaco I have a copy but haven't gotten to it. I am glad someone is read it at least.

[Decompiled @decompiled_dev]

@vesperdigital Acknowledging the downsides disarms critics. It becomes a judgment call, which he is very qualified to make. There aren't any specifics about "too much". All he needs is to show consideration of the risks.

He's been pro-stimulus and cheap money at BoC

https://t.co/E4frrBfB0v

[Decompiled @decompiled_dev]

@broadglow Chinese policy makers are blind. Acclerationists are ignorant and open source advocates are confused.

Its hard to take seriously when he has to resort to status attacks against every view point that's not his own.

Regulatory FUD as a strategy is saying the quite part out loud

[Decompiled @decompiled_dev]

@finbarrtimbers @natolambert @deanwball Moral hazard as a service

[Decompiled @decompiled_dev]

@deanwball @natolambert Anyone releasing the open weight model can put in their own training data and tilt it to favour their preferred ideology. Not all incentives are measurable in $

[Decompiled @decompiled_dev]

@deanwball > One probable outcome of an open-weight-model-dominant world is full AI communism

Linux is provided freely as a public good and is provided as digital public infrastructure. No state required.

Does this mean we are living in operating system communism?

[Decompiled @decompiled_dev]

@finn_hulse I'd be quoting The Way of Code by Rick Rubin all the time:

The Vibe Coder is above people, and no one feels oppressed. He stands ahead, and no one’s left behind.
Because he does not compete, he will not have competition.