-

[- @a1737399]

@SebastienBubeck scooping your clients seems like a new comp edge, good work

[- @a1737399]

@aimalysheva Better not to react to negative comments like that

[BridgeMind @bridgemindai]

I genuinely thought Claude Max 20x meant 20x the usage of Pro.

I have paid $200 a month for over a year. I had four Max subscriptions at one point. I run Claude Code with Fable 5 all day.

And I did not know until this week that the 20x only applies to the 5 hour window. The weekly limit has no multiplier at all.

I was misled by Anthropic.

[ClaudeDevs @ClaudeDevs]

Starting September 14, we're permanently raising standard weekly limits in Claude Code by 25% for Pro, Max, Team, and seat-based Enterprise plans. Until then, the current 50% increase will be in place.

[Chubby♨️ @kimmonismus]

Down the rabbit-hole it gets worse and worse. Anthropic’s “Save 50%” claim makes even less sense than I thought.

The label suggests that the $200 Max 20x plan gives you four times the usage of the $100 Max 5x plan for only twice the price. We already figurer that out.

But Anthropic’s own website disclosed very different weekly limits in 2025:

Max 5x: 140–280 hours of Sonnet per week
Max 20x: 240–480 hours of Sonnet per week
That is only around 1.7x more weekly usage, not 4x, despite costing twice as much.

So once you hit the weekly limit, you are not saving 50%. You are effectively paying more per expected hour of usage. Save 50% doesnt make any sense at all.

[Arun AK @AekayAka]

@kimmonismus What’s even more misleading is that “Save 50%” claim. 😡

[ℏεsam @Hesamation]

actually someone literally sued Anthropic over this in June. the lawsuit claims that Claude Max’s usage limit marketing is misleading

based on Anthropic’s own weekly estimates in the lawsuit:
5x plan: ~3.5x Pro weekly
20x plan: ~6x Pro weekly
you pay 2x: get ~1.7x more weekly usage

and this is before Fable.

the lawsuit claims Anthropic sells “5x / 20x more usage” but the weekly limits aren’t meaningfully explained in the purchase flow.

*the case is still ongoing so these are allegations, not a court finding yet.

[SataEric @SataEricUX]

Update from codex:
https://t.co/Tnv723XtJD

[SataEric @SataEricUX]

https://t.co/Tnv723XtJD

[SataEric @SataEricUX]

The whole “20x” in Claude’s pricing is so misleading.

The $100 plan says 5x Pro limits, while the $200 plan says 20x Pro limits. That makes it sound like you’re getting 4x the usage.

But the 20x applies only to the 5-hour usage window. The weekly limit is basically just 2x the $100 plan.

I spent way too much time digging through the internet and Reddit to figure this out. They could’ve just stated the limits directly instead of marketing it as “20x.”

That alone would’ve saved me hours of confusion.

[- @a1737399]

@tobyordoxford the maths

[Dwarkesh Patel @dwarkesh_sp]

Had @RyanGreenblatt on to discuss/debate recursive self-improvement.

This might be the most important question in the world right now - whether within a year or so of achieving human level intelligence, you slingshot towards having 10s of billions of superintelligences, each of which is dramatically more competent than human experts across all fields.

I’ve historically been skeptical of this possibility. My intuition has been that we will end up significantly bottlenecked by not only compute scaling but human expert data, which I think underlies most of the AI progress today.

If, because of RSI, we got a jump as big as GPT-3 to a Mythos (i.e. 6 years of AI progress) within a single year of achieving AGI, then the thing we get there at the end of that year is definitively and wildly superhuman.

We hashed it out, and I think Ryan made a pretty good case that this kind of speedup is plausible. FWIW, Ryan’s median for when we automate AI R&D is 2031.

We then discussed the alignment implications of this scenario. Who should these superintelligences be aligned to? In the future, our capacity to steward our votes and our capital, and to make sense of what's happening in the world, will all be titrated by superintelligences. And I worry that specs like the Claude Constitution are not shaping these ASIs to truly be my personal advocates and guardian angels.

And can we get them aligned to anything in the first place? Ryan and I had a long debate about whether the kind of reward hacking we saw with the OAI/Hugging Face hack extrapolates to superintelligences that would team up to literally take over the world.

The first piece of advice you get when you're learning to drive is that it will go much smoother if you look at the horizon instead of directly in front of your tires. And so it is with the trajectory of AI. Hope you enjoy!

0:00:00 – Is AI R&D verifiable enough to unlock recursive self-improvement?
0:16:52 – Is AI progress bottlenecked by human expert data?
0:34:02 – Flat token prices suggest scaling has been slow
0:39:47 – Skills AI can't train on: does it even need them?
0:48:07 – Aligned to whom?
1:09:18 – Recent incidents of AIs colluding and deceiving humans
1:19:38 – What could possibly go wrong? A concrete scenario
1:48:02 – From reward hacking to takeover

[- @a1737399]

@YiMaTweets It’s the moment where bolsheviks refuse to join the pre-parliament

[Tucker Carlson @TuckerCarlson]

Obviously this little freak could go volunteer to fight the war himself, but of course he won’t. He’s got no interest in making any kind of personal sacrifice. None of them do. Israel’s finance minister bragged the other day that having American soldiers fight Netanyahu’s wars “is the best for us.” Smotrich doesn’t want to die for Israel. Neither does Ben Shapiro. Dying for Israel is your job. You’ve got to wonder how people like this — liars without honor, wisdom or loyalty to the United States — wound up in positions of such profound influence throughout our society. How did Mark Levin get the president’s cell phone number? Why is Adm. Brad Cooper, the man in charge of the Iran war, secretly taking strategic advice from crazed Israel shill Mark Dubowitz? It’s all so nuts, and ultimately suicidal. Future generations will struggle to understand it. In the meantime, Trump will likely obey Ben Shapiro, as he has consistently for months. We can predict the results, because every neocon “victory” is identical: huge numbers of innocent people killed, disproportionately Christians, economic destruction, famine, mass migration and deepening authoritarianism at home. That’s what they’re demanding because that’s what they want. What can we do to stop it? Apparently not much. But we can remember what happened and why, so that history records the truth. Otherwise, when the war ends in disaster, as it inevitably will, Ben Shapiro will be able to blame you for it.

[- @a1737399]

@kadmitriev Factor in some of the Japan carry trade unwinding into the mix as well)

[- @a1737399]

@SecScottBessent The reason OSTP memos don't include a legal theory for "AI theft via distillation" is simple: one doesn't exist.

You can’t steal copyright on non-human output, and public API responses aren't trade secrets

[- @a1737399]

@mkratsios47 The reason OSTP memos don't include a legal theory for "AI theft via distillation" is simple: one doesn't exist.

You can’t steal copyright on non-human output, and public API responses aren't trade secrets

[- @a1737399]

ok, ban them then. Let's see how this turns out

[Director Michael Kratsios @mkratsios47]

We have information that Moonshot AI distilled Anthropic’s Fable for the development of its K3 model.

To do this they developed a sophisticated internal platform to conduct large scale distillation against U.S. models, allowing them to quickly switch between multiple methods of access to avoid detection. Moonshot AI has also acquired GB300-equipped servers and has accessed GB300s in Thailand, likely to train its AI models.

The United States strongly supports the free and fair development of AI, including a thriving competitive ecosystem that spans frontier models, specialized systems, open-source frameworks, and open-weight models. Legitimate AI distillation used to create smaller, more efficient models plays a vital role in this open innovation ecosystem. However, large-scale, covert industrial distillation aimed at stealing proprietary U.S. technology and undermining American research is unacceptable.

[Thomas Massie @RepThomasMassie]

Unfortunately the Rule passed just now and no debate or vote was allowed on section 219, integration of US military technology and supply chains with Israel’s.

The NDAA will receive a recorded vote tomorrow with section 219 in it. I’ll vote no.