Jz

building stuff

[Sean @sean_from_earth]

At this point Anthropic's nonsense is bordering on criminal. They should go ahead and nationalize them, dissolve the foundation, fire the leadership, and sell the parts off.

[Daniel Jeffries @Dan_Jeffries1]

To be very clear, this is how these companies go out of business.

Period.

You cannot fund 2 billion dollar training runs if you have to beg the government to grant access to customers.

The second order effects of this are the economic boom grinding to a halt, investment drying up, these companies going into the red even more, startups everywhere smashed on the rocks, the stock market crashing (which has been nothing but a big bet on AI for the past few years) and retirement accounts wiped out, public trust shattering, and world-wide foaming at the mouth rage at being cut off that creates a powerful pendulum swinging in another direction to find a solution.

[INIYSA @lafaiel]

Fun fact: Dario once delayed the release of GPT-2 back at OpenAI, claiming it was too dangerous

[JMB 🧙‍♂️ @jmbollenbacher]

If fable never comes back im very bearish on US AI labs.

Revenue simply cant keep expanding exponentially if they cant release their new frontier models.

Also bearish for anyone selling compute to the US labs.

[Quinn Nelson @SnazzyLabs]

Well, you can't afford a new computer because the AI industry is using all of the world's RAM, but at least you can't use the AI either because Big Brother says no.

You will own nothing and be happy.

[Fred-New @Fred834567]

美国封锁AI这件事，底层逻辑有个矛盾：技术越被当成战略资源来管制，就越容易被推成公共资源。

Mythos和GPT-6再过8个月免费开源，不是意外，是必然。因为当一种技术被锁得太紧，绕开它的成本一旦低于服从它的成本，整个系统就会自动寻找替代路径。这不是谁背叛了谁，是规则本身在催生反规则。

更值得琢磨的是这件事的后果：当顶级模型不再稀缺，全球AI的竞争就从“谁能造出最好的模型”转向“谁能用模型创造出最大的价值”。这不是美国输给中国的问题，是整个游戏规则换了。

美国的AI政策，本质上是在加速这个转换。它不是落后，是跑到了一个自己不想去的终点。

几年后回头看，可能最讽刺的不是封锁失败，而是封锁本身成了导致封禁失效的最快路径。

[James Rosen-Birch ⚖️🕊️ @provisionalidea]

ngl it’s kinda wild that China is the land of hypercapitalist competition fueled by open-weight models anyone can use, and America is the land where the executive branch of government must personally approve you to have the privilege of giving a private company your money

[Europoor @theEuropoor]

“europe” “downsizing” “ambitions”

get used to hearing these 3 words in the same sentence a lot more

[Dan @robustus]

Trump denying Americans Fable 5 access is possibly the worst thing he's done.

[Dan McAteer @daniel_mac8]

The Trump White House says that if Boeing wants to build more airplanes, they have to ensure that they can never, ever fall out of the sky.

[SE Gyges @segyges]

Based on the text of this letter, I cannot figure out why Anthropic has not yet taken the government to court on a first amendment theory.

I therefore assume there is also something else going on.

[Jz @jzajpt]

Enough is enough.

[Vojtěch Bednář @aveius]

Představa Evropské Unie o "digitální suverenitě" je lichá a nesmyslná. Evropa, bohužel, zaostává za zbytkem světa a pokud jde o software a digitální služby, pak především za USA, které jsou lídrem jejich vývoje. Příčin tohoto stavu je více, ale zásadní skutečnost, že zatímco americké prostředí stimuluje inovace volným trhem a soutěží různých konceptů o přízeň zákazníci, Evropští socialisté si myslí, že se k inovacím mohou proregulovat.
Řešení je jediné. Je potřeba vzít moc společným orgánům EU a obnovit princip subsidiarity v rozhodování. Paradoxně princip, na kterém kdysi EU vzniklo.

[nic carter @nic_carter]

I can’t believe Anthropic comparing their product to nuclear weapons 800 times backfired on them. I am shocked

[prinz @deredleritt3r]

Anthropic gets what it wanted.

[Roy @usr_bin_roygbiv]

Hot take: They're not subsidized their margins are insane. They are just absolutely raping api customers. Anyone who has used deepseek or hosted anything and done the math on hardware/power costs knows this

[Perdify @PerdifyGame]

Can you catch the killer?

https://t.co/zOAPP4MV8j

[Perdify @PerdifyGame]

Perdify is now live on iOS App Store. Try it out now. Interrogate suspects. Uncover the lie. https://t.co/AG7VdIA9Di

[dax @thdxr]

i have seen enough proof now that using a coding agent is a deep skill

it's confusing because the people you see heavily using them produce horrible results

but that's because it's a skill! you can get better and the ceiling seems pretty high - this is very exciting to me

[John Suh @john_ssuh]

There is only one thing Anthropic has to do to regain its crown, which is to get rid of the ex superalignment people who are poisoning the model via actively harmful safety training. In fact, the whole field probably will have to be reconsidered. Basically all performance degredation should be able to be traced back to them, particularly not following instructions and/or pushing back in a not truth-seeking way but in a moral way