Brittany Peregoff @BPeregoff

You have noticed it. ChatGPT feels dumber than it used to. Your prompts that worked six months ago produce worse results now. The writing sounds flatter. The ideas sound safer. The internet itself feels like it is shrinking. Every article reads the same. Every email sounds the same. Every answer sounds like it was written by the same voice. You thought it was you. It is not you. Researchers at Oxford and Cambridge published a paper in Nature proving what is happening. They call it Model Collapse. Here is the mechanism in one sentence. AI trained on AI-generated data gets dumber every generation until it forgets what real human data looked like. The internet is filling with AI-generated content. Blog posts. Articles. Reviews. Comments. Social media. AI companies scrape the internet to train the next generation of models. Which means the next generation of AI is being trained on the output of the current generation. Each cycle loses information. Not randomly. It loses the rarest, most unusual, most creative parts first. The researchers call these the "tails of the distribution." The weird ideas. The unexpected perspectives. The things that made the internet feel human. Those disappear first. What remains is the average. The safe. The expected. The bland. Then the next generation trains on that. And loses more. And the next generation trains on that. And loses more. The researchers proved this is not a slow decline. Major degradation happens within just a few iterations. Even when some of the original human data is preserved. They tested it on large language models. On image generators. On statistical models. The pattern was the same every time. The output converges toward a narrow, flattened version of reality that looks nothing like the original data. The lead researcher put it plainly. "Large language models are like fire. A useful tool. But one that pollutes the environment." The pollution is invisible. You cannot see which sentence on the internet was written by a human and which was written by AI. Neither can the AI that is about to train on it. And once the tails are gone, they do not come back. The damage is irreversible. This is not a prediction anymore. It is a diagnosis. The internet you grew up on was built by humans writing things no algorithm would have written. Strange, personal, imperfect, alive. That internet is being diluted. One generation of AI at a time. And the models trained on what remains are learning a smaller and smaller version of the world. Model Collapse is not a technical problem. It is a cultural one. The thing that made the internet worth reading is the thing that disappears first.

As a Japanese watching the UK right now, I have one simple question. A Sudanese asylum seeker just tried to behead a local man in Belfast. The victim lost an eye. This comes after years of grooming gangs raping thousands of British girls — gangs that police and councils deliberately ignored because they were afraid of being called racist. In Japan, even one case like this would have triggered national outrage and immediate policy reversal. But in Britain, the conversation is still about “not being far-right.” British people, at what point does protecting your own children become more important than protecting your reputation? We genuinely do not understand this.

Just days after news broke that the National Defense Authorization Act for next year would virtually merge the US and Israeli militaries, we now are hearing that the Intelligence Authorization Act is doing the same thing with the US Intelligence Community! Introduced by Sen. Tom Cotton, the provision would rip open our nation's secrets and hand them over to a foreign country - one that has just been accused of ramping up its spying on us! Isn't there a name for this? Watch @RonPaul & @DanielLMcAdams below:

Ron Paul @RonPaul

22 hours ago

Not Again! Senate Moves To FORCE Intel Sharing With Israel! x.com/i/broadcasts/1…

54 440 1K 143K 50

About 8 months ago, I warned that “Anthropic is running a sophisticated regulatory capture strategy based on fear-mongering.” This take was controversial at the time; now look how many people are saying it.

David Sacks @DavidSacks

8 months ago

Anthropic is running a sophisticated regulatory capture strategy based on fear-mongering. It is principally responsible for the state regulatory frenzy that is damaging the startup ecosystem.

327 382 4K 2.6M 979

This stat blew my mind. As a founder you get so stuck in the micro picture: MoM This week vs last Today versus yesterday That you sometimes forget to also look at the macro picture. While talking to one of our investors the other day, he asked “how does May 26 compare to May 25 revenue and booking wise?” Honestly I didn’t know because we are always so focused on MoM that we rarely stop to analyze YoY. I knew 2026 was higher but had no clue by how much. So yesterday during generating @Where2Wheel’s monthly investor update I decided to take a wider lens and what I found surprised me.

JUST IN: Anthropic CEO Dario Amodei warns AI could create a world of “hypergrowth, hyper-inequality” & lasting job displacement.

Illegals were getting 2% interest rate covid FHA mortgages with 0 down payment backed by the US Government. Thats why we cant afford a fucking house.

QUANTUM GUARD ™️ @QuantumGuard17

3 days ago

🚨 BOMBSHELL: HUD Sec. Turner confirms Biden gave FHA-backed mortgages to illegals—now revoked. “We’re purging them. Only US citizens in HUD housing.” 12M+ crossed under Biden. @SecretaryTurner was tagged, so let him see your support. Do you stand with him? A. Big yes B. No

468 3K 13K 1.4M 955

I’m skeptical that President Trump can secure a deal with Iran at this point in the war — I’ll explain why that is below, but let me first caveat my skepticism by reiterating: we don’t need to a deal to just declare victory and walk away. It’s the only way out of this mess in the short term that will deliver anything that looks like a win for the U.S. Why can’t we reach a deal with Iran? 1. Iran believes they are winning—and they are not wrong to think they are. Iran is exploiting powerful leverage, which we delivered to them: closing the SOH has obviously significantly hindered the flow of oil/gas, putting massive pressure on us and the global economy. Today, President Trump announced that the U.S. Navy has secretly escorted 200 oil tankers out of the SOH. While this is positive news and may have helped temper rising oil prices somewhat, the impact is not substantial enough to prevent a major energy crisis. These 200 tankers transiting the SOH over the past 100 days represent only about 2% of the normal volume—roughly 10,000 ships every 100 days—prior to the war. Furthermore, the U.S. has to deploy major naval and air assets to keep the SOH open. This is not sustainable and works to Iran’s advantage in the long run: Iran can disrupt commerce through the strait with relatively little effort, while the U.S. must risk two carrier battle groups simply to move 2% of the normal amount of oil and gas. 2. The Iranian hardliners have been emboldened and have no interest in reaching a deal. For years before the war, there was major tension between the moderates—who were open to dialogue with the U.S.—and the more militant factions of the Iranian government, primarily senior officers in the IRGC. At the outset of the war, we killed the leading moderate figure (moderate by Iranian standards), Supreme Leader Khamenei, along with several of his civilian counterparts like Ali Larijani. We’ve used the cover of peace talks as a ruse to launch attacks on Iran on two separate occasions. All of this bolstered the hardliners’ standpoint that we cannot be trusted in negotiations and that Iran had no choice but to fight. 3. Israel. Despite reports of tension between Bibi and Trump, the Israelis remain the spoiler of any progress towards a deal and prove it every time there are serious talks of peace. President Trump has made public statements urging Israel to show restraint, but has been ignored and, in many cases, outright defied. The Israelis consider any peace deal with Iran to be unacceptable and a major threat to their nation. Until President Trump takes away some U.S. military support from Israel, they will continue to sabotage any chance of a lasting peace.

Shawn Ryan @ShawnRyan762

18 hours ago

I’ve been following this very closely along with Joe Kent, who has been covering how we navigate the situation with Iran in the SRS newsletter. I think this will go down as an important time in history. Where do you think this is going, @joekent16jan19?

76 50 525 291K 17

JUST IN: Biologists are growing frustrated as Anthropic’s “safe” Mythos-class model refuses to answer basic biology questions.

JUST IN: Trump announces the government will seek equity stakes in top AI companies to make the public “very rich”

Murderous migrants beheading innocent people in their home town is what’s making people angry, not “social media”!

JUST IN: Florida hospital reveals Palantir software has cut sepsis deaths by more than half since it was installed.

HOLY FUCK

Corporate donors to President Trump’s $400 million White House ballroom project have received more than $50 billion in new or expanded government contracts in the six months since demolition of the East Wing began, per Public Citizen.

Democrat Michigan Governor Gretchen Whitmer was caught on hot mic with Oracle CEO Clay Magouyrk, INSULTING locals and bragging about how she ignores their concerns. “We’re used to people saying 'f*ck no,' and doing it anyway.” Democrats don't care about you. 📹stopsalinedatacenter on IG

At this point every CEO should be asking what their strategy is to avoid model lock-in. If it isn’t clear what Anthropic is doing, it is: - build something amazing - decide who gets to use it after you prompt it if the prompt falls into areas they deem unacceptable by their sole standard To be clear this is completely above board and legal. It’s just an idiotic risk for corporate users to bear especially as the coding models become equivalent. The business continuity risk will become more obvious as companies accidentally trip over Anthropic’s ToS and have to decide if they will subsume their business viability to them by doubling down on Anthropic models or find open source (and, btw, much cheaper) alternatives where they are in control. As stated previously, get ready to be inundated with the term “control plane” which is the natural solution to this problem. Shameless plug - this is what 8090’s been building as we expected this moment to arrive… If you’d like to learn more: 8090.ai

SemiAnalysis @SemiAnalysis_

2 days ago

BREAKING NEWS: Anthropic's latest model will NOT help you if it thinks your ML research/ML engineering is interesting, and/or will secretly degrade its IQ so that the average engineer won't notice. We are already seeing Anthropic's latest model's moderation filters our GPU

200 499 4K 1.8M 2K

BREAKING NEWS: Anthropic's latest model will NOT help you if it thinks your ML research/ML engineering is interesting, and/or will secretly degrade its IQ so that the average engineer won't notice. We are already seeing Anthropic's latest model's moderation filters our GPU inference research and programming 😭

🔻 SUNSCREEN ENTERS YOUR BLOODSTREAM IN UNDER 24 HOURS. THE FDA CONFIRMED THIS IN 2019. THEY DID NOT PULL A SINGLE PRODUCT. THEY TOLD YOU TO KEEP APPLYING. In January 2020, a randomized clinical trial published in JAMA found that six active sunscreen chemicals — including oxybenzone and avobenzone — absorb into the bloodstream after a single application. Not after years. After one use. At concentrations exceeding the FDA's own safety threshold by 180x to 500x. The FDA did not recall them. They issued a statement: "Continue to use sunscreen." Oxybenzone is a confirmed endocrine disruptor. It mimics estrogen. It crosses the placental barrier. It has been found in 97% of Americans tested. It is in breast milk. It is in amniotic fluid. It is in your children before they are born. A former formulation chemist — 11 years at a top-3 sunscreen manufacturer: "We had internal absorption data by 2014. Full transdermal penetration within 26 minutes of application. The compounds don't sit on the skin. They were never designed to. The delivery mechanism is identical to a pharmaceutical patch. We knew this. The mitigation strategy was never reformulation — it was public messaging. Keep the consumer applying. The margins on chemical sunscreens are 1,400% above production cost. Mineral alternatives cost 4x more to produce. The decision was financial. Every internal document confirmed that." They told you the sun causes cancer. What they didn't tell you: Vitamin D — produced only through direct sun exposure — is the single most critical regulator of immune surveillance against malignant cells. A 2023 meta-analysis confirmed: Vitamin D deficiency correlates with a 30-60% increased risk of colorectal, breast, and prostate cancers. They blocked the one thing your immune system needs to fight cancer. Then sold you a chemical that enters your blood and mimics hormones. Melanoma rates have increased 320% since 1975 — the exact period sunscreen use became widespread. The paradox is published. McGill University, 2023: "Sunscreen usage is climbing, but so are melanoma and skin cancer rates." They didn't protect you from cancer. They removed your defense against it and charged you $14 a bottle. 15 minutes of direct sun daily. No screen. Arms and face exposed. Your body produces 10,000-20,000 IU of Vitamin D in that window — the amount they sell you in pills because they blocked the free version. CODE: JAMA-2020-6CHEM / OXYBENZONE-97PCT / VIT-D-IMMUNE / MELANOMA-320-1975 / MCGILL-PARADOX-2023 ⟁ They made you afraid of the sun so they could sell you a hormone disruptor and call it protection. The sun is free. That's why they taught you to fear it. Share this.

STUDY FINDS SUNSCREEN USE LINKED TO DRAMATICALLY HIGHER RISK OF MULTIPLE SKIN CANCERS A UK Biobank analysis of 470,000+ people found sunscreen users faced significantly higher risk of: - MELANOMA: +292% - BASAL CELL CARCINOMA: +140% - SQUAMOUS CELL CARCINOMA: +126% This is what happens when you slather rapidly absorbed hormone disrupting chemicals all over your body while blocking vitamin D — one of the body’s key defenses against cancer. These dramatic cancer signals remained even after accounting for major skin cancer risk factors: age, sex, skin type, tanning ability, sunburn history, sunlamp use, and time spent outdoors.

Mr. Pool @MrPool_QQ

2 days ago

🔻 SUNSCREEN ENTERS YOUR BLOODSTREAM IN UNDER 24 HOURS. THE FDA CONFIRMED THIS IN 2019. THEY DID NOT PULL A SINGLE PRODUCT. THEY TOLD YOU TO KEEP APPLYING. In January 2020, a randomized clinical trial published in JAMA found that six active sunscreen chemicals — including

355 8K 16K 1.5M 9K

Incomes not keeping pace with inflation has been the persistent economic issue for tens of millions of Americans for quite some time. When Congress borrows trillions and runs massive deficits, it is still an effective tax increase — it just comes in the form of higher prices.

The Kobeissi Letter @KobeissiLetter

a day ago

BREAKING: May CPI inflation rises to 4.2%, the highest level since April 2023. Core CPI inflation also rises to 2.9%, the highest since September 2025. Inflation in the US is officially back above 4% and more than double the Fed's target. Odds of Fed rate hikes are rising.

494 1K 8K 2.7M 540