Anthropic's Fable Backlash, Nationalizing AI, Inflation Heats Up & California's Broken Elections
Episode
102 min
Read time
3 min
Topics
Career Growth, Productivity, Personal Finance
AI-Generated Summary
Key Takeaways
- ✓Anthropic's Hidden Downgrade System: Fable 5 silently degrades users to a lesser model when its systems classify their prompts as frontier AI research, chip design, or sensitive science — without disclosure, while still charging full price. Anthropic has since committed to notifying users when downgraded, but the classification criteria remain broad enough to flag routine queries about mitochondria or GLP-1 cancer research, making enterprise reliance on the platform a single-point-of-failure risk.
- ✓Regulatory Capture Playbook: Anthropic's simultaneous rollout of restrictive model policies and Dario Amodei's public call for an FAA/FDA-style AI regulatory agency follows a recognizable pattern: incumbents invite regulation they help design, then use compliance costs to disadvantage open-source competitors who cannot be regulated the same way. Enterprises evaluating AI strategy should treat any vendor lobbying for model licensing regimes as a signal that open-source alternatives may face legislated suppression within 24–36 months.
- ✓Chinese Open-Source Models Fill the Gap: When closed U.S. models restrict genomic, biotech, or materials science research, companies default to the best available open-source alternatives — which are currently Chinese models. Freeberg's agricultural genomics firm is already migrating to locally-run open-source tools after Fable 5 blocked standard RNA guide design and genetic construct work. U.S. AI safety restrictions are functionally transferring competitive advantage in biotech and materials science to Chinese model providers.
- ✓Compute Capital Moat at $100B Per Gigawatt: Building one gigawatt of AI data center capacity now costs approximately $100 billion, a 20x increase from roughly $4–5 billion just a few years ago. Chamath holds 2,000 acres of zoned, approved land in Arizona capable of supporting two gigawatts and is bidding on additional capacity, concluding that without redirecting large-scale compute toward open-source models, the open-source ecosystem will remain marginal — most global GPU megawatts still flow to closed frontier labs.
- ✓Bernie Sanders' 50% AI Equity Seizure — The Steel Man: Sanders' proposed American AI Sovereign Wealth Fund Act calls for a one-time 50% stock seizure from OpenAI, Anthropic, and XAI, with shares granting public voting rights. The politically durable argument: AI companies trained on humanity's freely contributed knowledge, publicly forecast 50% entry-level job displacement, and now gatekeep outputs. Sacks argues the more defensible alternative is voluntary public participation structures, while Freeberg proposes reforming Social Security's $4 trillion treasury certificate into an equity-holding sovereign wealth fund.
What It Covers
Anthropic's Claude Fable 5 release triggers developer backlash over mandatory 30-day prompt retention and undisclosed model degradation for AI researchers, prompting debate on regulatory capture, open-source AI viability, Bernie Sanders' proposed 50% equity seizure of AI companies, May CPI hitting 4.2% year-over-year, and statistical anomalies in the Los Angeles mayoral primary ballot counts.
Key Questions Answered
- •Anthropic's Hidden Downgrade System: Fable 5 silently degrades users to a lesser model when its systems classify their prompts as frontier AI research, chip design, or sensitive science — without disclosure, while still charging full price. Anthropic has since committed to notifying users when downgraded, but the classification criteria remain broad enough to flag routine queries about mitochondria or GLP-1 cancer research, making enterprise reliance on the platform a single-point-of-failure risk.
- •Regulatory Capture Playbook: Anthropic's simultaneous rollout of restrictive model policies and Dario Amodei's public call for an FAA/FDA-style AI regulatory agency follows a recognizable pattern: incumbents invite regulation they help design, then use compliance costs to disadvantage open-source competitors who cannot be regulated the same way. Enterprises evaluating AI strategy should treat any vendor lobbying for model licensing regimes as a signal that open-source alternatives may face legislated suppression within 24–36 months.
- •Chinese Open-Source Models Fill the Gap: When closed U.S. models restrict genomic, biotech, or materials science research, companies default to the best available open-source alternatives — which are currently Chinese models. Freeberg's agricultural genomics firm is already migrating to locally-run open-source tools after Fable 5 blocked standard RNA guide design and genetic construct work. U.S. AI safety restrictions are functionally transferring competitive advantage in biotech and materials science to Chinese model providers.
- •Compute Capital Moat at $100B Per Gigawatt: Building one gigawatt of AI data center capacity now costs approximately $100 billion, a 20x increase from roughly $4–5 billion just a few years ago. Chamath holds 2,000 acres of zoned, approved land in Arizona capable of supporting two gigawatts and is bidding on additional capacity, concluding that without redirecting large-scale compute toward open-source models, the open-source ecosystem will remain marginal — most global GPU megawatts still flow to closed frontier labs.
- •Bernie Sanders' 50% AI Equity Seizure — The Steel Man: Sanders' proposed American AI Sovereign Wealth Fund Act calls for a one-time 50% stock seizure from OpenAI, Anthropic, and XAI, with shares granting public voting rights. The politically durable argument: AI companies trained on humanity's freely contributed knowledge, publicly forecast 50% entry-level job displacement, and now gatekeep outputs. Sacks argues the more defensible alternative is voluntary public participation structures, while Freeberg proposes reforming Social Security's $4 trillion treasury certificate into an equity-holding sovereign wealth fund.
- •AI and Job Creation — The Revenue-Side Argument: Contrary to Dario Amodei's forecast of 50% entry-level knowledge worker displacement within one to five years, ground-level data shows AI driving hiring increases, not cuts. The May jobs report showed 172,000 new positions, software developer roles at a three-year high up 15% year-over-year, and 4.3% unemployment. The mechanism: AI multiplies individual engineer output 100x to 1,000x, enabling companies to build more products and expand revenue rather than simply cutting headcount on the cost side.
- •California Ballot Statistics and Structural Vulnerability: In the LA mayoral primary, Spencer Pratt led in-person election-day voting at 35% versus Karen Bass at 29% and Nithya Raman at 26%. Among mail-in ballots arriving after election day, Raman's share surged 80% relative to pre-election mail-ins while Pratt's share dropped by one-third. California's combination of universal ballot mailing to 23–24 million registered voters, unlimited ballot harvesting under AB 1921, no voter ID requirement for registration, and 40%-threshold signature matching creates structural conditions the hosts argue enable systematic vote manipulation regardless of legal status.
Notable Moment
During the live recording, Jason Calacanis tested Fable 5's content restrictions in real time and was visibly downgraded mid-conversation — first for asking about fertilizer bomb regulations, then for inquiring about nuclear bomb components. The model's internal reasoning log, visible to the user, explicitly identified him as a venture capitalist and podcaster before restricting his access.
Episode Transcript
Alright, everybody. Welcome back to the number one podcast in the world. It's your favorite podcast. It's your podcaster's favorite podcast. It's the podcast that everybody likes to rip off for 50 episodes and then they quit. But the all in podcast is not quitting. We're doubling down with the original quartet. Everybody's here for a big, big week, lots of debates, and we gotta start with Anthropic. Anthropic released a mythos level model with some interesting guardrails and, it seems as though Dario is back to blogging. Okay. The model's called Fable five? No idea why they're calling it that. They released on Tuesday and it tops every benchmark nearly every benchmark. The tokens, however, cost twice as much as Opus four 0.8, but it should use less tokens overall because it's theoretically that much better. Unclear if this is going to alleviate some of the hand wringing around the cost of tokens. But you'll remember in April, Anthropic famously did not release ETHOS because publicly, it had some strong hacking capabilities, so they gave it to some folks and we had, the CEO of, Nakesh of Palo Alto Networks who said, hey, Methos was the real deal and that they used it to seal up some vulnerabilities inside their shop. Obviously this new model is sensitive, to topics like bioweapons, hacking, all that stuff are blocked in it, but they got a big developer backlash, and then I'll get your feedback after I explain this. While using Fable, Anthropic stores all the prompt data you enter for at least thirty days. And so that's a bit of a privacy issue. And if Fable five detects you're doing frontier AI research, in other words, using their model to make a better model, competitive model. They were downgrading you, but they weren't telling you. And this was buried deep in a 319 page document. So many kerfuffles, freak outs going on on x over this, but they've since walked it back a bit, quote unquote, in Wired. We're changing Fable five safeguards for Frontier LLM development to make them more visible. I'm gonna stop there, Chamath. What's your take on anthropic and these extraordinary models they keep dropping and how they're handling the release of them. Are they being thoughtful? Are they being dramatic and drama queens? A little bit of both. Where do you stand on it, you know, after getting Nikesh's feedback and seeing the latest model come out? Have you played with it? What's going on at eighty, ninety in the software factory in terms of benchmarking it as well? It's a really incredible model. And so kudos to these guys for continuing to push the boundaries of the closed Frontier Lab models. They're firing on all cylinders. I think that it creates a pretty obvious risk now. And I think that obvious risk is twofold. One is Anthropic has essentially shown their hand, which is that they will increasingly take in prompts, evaluate the prompts and …
Get the full transcript (19,014 words) + summary by email — free
One-time email with the complete transcript and AI summary of this episode. No account needed.
One email, no spam. We’ll also show you what SignalCast does.
Browse all All-In with Chamath, Jason, Sacks & Friedberg transcripts →
You just read a 3-minute summary of a 99-minute episode.
Get All-In with Chamath, Jason, Sacks & Friedberg summarized like this every Monday — plus up to 2 more podcasts, free.
Pick Your Podcasts — FreeKeep Reading
More from All-In with Chamath, Jason, Sacks & Friedberg
AI Kills Everybody or Doomer Psyop? OpenAI's Math Breakthrough, Nike's $200B Collapse
Sep 11 · 95 min
The AI Breakdown
Why Fable 5.1 Is Worth the Upgrade
Sep 2
More from All-In with Chamath, Jason, Sacks & Friedberg
GPT-6 Hits AGI? Tech Euphoria 2.0, SF Mansion Shortage, NYC Bans AI in Schools & Venezuela Oil Deal
Sep 4 · 91 min
The AI Breakdown
Where Claude Opus 5 Fits in Your Model Rotation
Jul 27
More from All-In with Chamath, Jason, Sacks & Friedberg
We summarize every new episode. Want them in your inbox?
AI Kills Everybody or Doomer Psyop? OpenAI's Math Breakthrough, Nike's $200B Collapse
GPT-6 Hits AGI? Tech Euphoria 2.0, SF Mansion Shortage, NYC Bans AI in Schools & Venezuela Oil Deal
Nvidia's Historic Quarter, SaaS Comeback, Bessent vs Druck, America's Debt Crisis, Cancer Vaccine
Eric Weinstein: The State of American Science, Breakthrough Coverups, and the Danger of Physics
Michael Kratsios: Trump's Science Agenda, Anti-Science Claims, Fauci's Damage, DEI & China
Similar Episodes
Related episodes from other podcasts
The AI Breakdown
Sep 2
Why Fable 5.1 Is Worth the Upgrade
The AI Breakdown
Jul 27
Where Claude Opus 5 Fits in Your Model Rotation
Odd Lots
Jul 20
The Creator of Claude Code on The Hottest Piece of Software in the World
The AI Breakdown
Jul 1
Fable is Back: Here's What You Should Try First
The AI Breakdown
Jun 22
Why AI Users Are Raving About GLM 5.2
Explore Related Topics
This podcast is featured in Best Tech Podcasts (2026) — ranked and reviewed with AI summaries.
You're clearly into All-In with Chamath, Jason, Sacks & Friedberg.
Every Monday, we deliver AI summaries of the latest episodes from All-In with Chamath, Jason, Sacks & Friedberg and 192+ other podcasts. Free for one show.
Start My Monday DigestNo credit card · Unsubscribe anytime