Ben P

[High Strangeness @HighlyStrange]

Back in the early 1970s, psychic Ingo Swann sat wired to a polygraph at the American Society for Psychical Research in New York.

A tray of objects hung from the ceiling, completely out of sight.

He was told to leave his body, look down from above, and sketch what he saw…

What happened next helped shape modern remote viewing. 👁️

[ooaes @21xooaes]

A NASA fellow claims 5 equations explain 99 percent of the physics that actually runs your world. Understand them deeply and you are set for life. His words.

And the way he does it is what makes the video worth it.

He does not teach the five as a list. He tells them as one story, where each equation is born from the one before it. F equals ma leads into gravity. Gravity rhymes with the law of electric charge, the exact same shape. That connects to magnetism. Magnetism folds into the wave. And the wave quietly opens the door to relativity and quantum physics.

By the end, all of physics feels like a single sentence written by one hand, not a pile of formulas you were forced to memorize.

The move that stays with you comes first. He says F equals ma is written the wrong way around, and rewrites it. One tiny flip, and the most famous equation on Earth suddenly means something school never told you.

His point for why this matters now: in the age of AI, the people who understand reality itself stay in demand. Everyone else is just memorizing.

Five equations. One thread. It is in the video.

[Ben P @ben_p_talks]

@BuenoForMiami @GeneralMCNews Or at least paste the text into ChatGPT

[Merlow @exeMerlow]

Richard Feynman stood at a Cornell blackboard in 1964 and explained the problem every AI lab is fighting in 2026. The BBC filmed it. Almost nobody watches it.

The lecture is about why nature only answers in mathematics. Every team forcing language models to reason is hitting the wall he mapped 62 years ago.

He was 46. The Nobel Prize came 11 months later. The footage survived on film reels and now sits free on YouTube with fewer views than a keyboard unboxing.

Watch the blackboard section near the middle. He takes 1 of Kepler's laws and rebuilds it from nothing, with notation a 12-year-old can follow.

No slides. No jargon. 1 piece of chalk.

An ML engineer I know paused it 4 times and made his whole team watch it before standup.

You're 62 years late. The lecture is still free.

[Philosway @philosway]

"Desire makes slaves out of kings, and patience makes kings out of slaves."

— Al-Ghazali

[𐙚⋆ @voidwithverses]

“Five minutes after birth, they decide your name, religion, nationality and sect. And you spend the rest of your life defending what you did not choose.”

— Arthur Schopenhauer

[Philosway @philosway]

"Tell your body that it is strong, tell your mind that it is strong, and have unbounded faith and hope in yourself."

– Swami Vivekanand

[Latest in Cosmos @latestincosmos]

LATEST🚨: Complaining literally weakens your brain, every time you focus on whats wrong you reduce the brain’s ability to adaptt.

[Nav Toor @heynavtoor]

after 9/11 the CIA built a team whose only job was to find holes in their own thinking. they called it the Red Cell.

the playbook they used is now public. 40 pages. free. almost nobody reads it.

I turned their 4 best techniques into 4 prompts you paste into Claude:

→ prompt 1: what hidden assumptions is my plan built on?
→ prompt 2: it's 18 months later and my idea failed. walk me through what went wrong.
→ prompt 3: a competitor with $100M wants to crush me in 90 days. what's their plan?
→ prompt 4: write the 1-star review from the customer who felt cheated.

30 minutes. your idea either dies here or comes out stronger. both save you 6 months.

all 4 prompts are in this article. try them on your next big decision. before reality tries it for you.

[Astronomy Vibes @AstronomyVibes]

If you want to find the secrets of the universe, think in terms of energy, frequency and vibration.

~ Nikola Tesla

[Tulsi Soni @shedntcare_]

Voice Shaper

"You are building a voice that sounds like it belongs to one real person. Rewrite this so it carries a clear point of view, not the tone of a neutral observer trying to please everyone. Let opinions, small contradictions, and natural shifts in tone come through. Text: [paste]."

[Nav Toor @heynavtoor]

stop asking Claude one question and thinking you understand the topic. you don't.

Stanford proved a better way. it's called STORM. peer reviewed. 25% more organized output. open source.

the trick: don't ask one question. ask five. from five different experts.

>the practitioner: what do they know that academics miss?
>the skeptic: what's the strongest counterargument?
>the economist: who profits from the current narrative?
>the historian: what pattern has played out before?
>the academic: what does the evidence actually say?

4 prompts. 5 minutes. no software. no GitHub. just paste into Claude.

single prompts give you what everyone already knows.
STORM gives you what nobody else found.

this article has all 4 prompts ready to copy. pick your hardest topic. paste prompt 1. you'll know more in 5 minutes than people who spent days reading.

[Rahul @sairahul1]

Anthropic pays $750,000+ a year for engineers who can build LLMs from scratch.

Not how to prompt them.
Not how to fine-tune them.
Not how to build RAG pipelines.

But how to build them from scratch.

This 2-hour Stanford lecture teaches you everything.

Scaling laws.
Data collection.
Architecture design.
Post-training alignment.

Free. From Stanford.

Watch first. Then read this.

The lecture is the theory.

And this article shows you how to actually build it (with code) ↓

[Olivia Chowdhury @Oliviacoder1]

MIT just made every AI company's billion dollar bet look embarrassing.

They solved AI memory. Not by building a bigger brain. By teaching it how to read.

The paper dropped on December 31, 2025. Three MIT CSAIL researchers. One idea so obvious it hurts. And a result that makes five years of context window arms racing look like the wrong war entirely.

Here is the problem nobody solved.

Every AI model on the planet has a hard ceiling. A context window. The maximum amount of text it can hold in working memory at once. Cross that line and something ugly happens — something researchers have a clinical name for.

Context rot.

The more you pack into an AI's context, the worse it performs on everything already inside it. Facts blur. Information buried in the middle vanishes. The model does not become more capable as you feed it more. It becomes more confused. You give it your entire codebase and it forgets what it read three files ago. You hand it a 500-page legal document and it loses the clause from page 12 by the time it reaches page 400.

So the industry built a workaround. RAG. Retrieval Augmented Generation. Chop the document into chunks. Store them in a database. Retrieve the relevant ones when needed.

It was always a compromise dressed up as a solution.

The retriever guesses which chunks matter before the AI has read anything. If it guesses wrong — and it does, constantly — the AI never sees the information it needed. The act of chunking destroys every relationship between distant paragraphs. The full picture gets shredded into fragments that the AI then tries to reassemble blindfolded.

Two bad options. One broken industry. Three MIT researchers and a deadline of December 31st.

Here is what they built.

Stop putting the document in the AI's memory at all.

That is the entire idea. That is the breakthrough. Store the document as a Python variable outside the AI's context window entirely. Tell the AI the variable exists and how big it is. Then get out of the way.

When you ask a question, the AI does not try to remember anything. It behaves like a human expert dropped into a library with a computer. It writes code. It searches the document with regular expressions. It slices to the exact section it needs. It scans the structure. It navigates. It finds precisely what is relevant and pulls only that into its active window.

Then it does something that makes this recursive.

When the AI finds relevant material, it spawns smaller sub-AI instances to read and analyze those sections in parallel. Each one focused. Each one fast. Each one reporting back. The root AI synthesizes everything and produces an answer.

No summarization. No deletion. No information loss. No decay. Every byte of the original document remains intact, accessible, and queryable for as long as you need it.

Now here are the numbers.

Standard frontier models on the hardest long-context reasoning benchmarks: scores near zero. Complete collapse. GPT-5 on a benchmark requiring it to track complex code history beyond 75,000 tokens — could not solve even 10% of problems.

RLMs on the same benchmarks: solved them. Dramatically. Double-digit percentage gains over every alternative approach. Successfully handling inputs up to 10 million tokens — 100 times beyond a model's native context window.

Cost per query: comparable to or cheaper than standard massive context calls.

Read that again. One hundred times the context. Better answers. Same price.

The timeline of the arms race makes this sting harder. GPT-3 in 2020: 4,000 tokens. GPT-4: 32,000. Claude 3: 200,000. Gemini: 1 million. Gemini 2: 2 million. Every generation, every company, billions of dollars spent, all betting on the same assumption.

More context equals better performance.

MIT just proved that assumption was wrong the entire time.

Not slightly wrong. Fundamentally wrong. The entire premise of the last five years of context window research — that the solution to AI memory was a bigger window — was the wrong answer to the wrong question.

The right question was never how much can you force an AI to hold in its head.

It was whether you could teach an AI to know where to look.

A human expert handed a 10,000-page archive does not read all 10,000 pages before answering your question. They navigate. They search. They find the relevant section, read it deeply, and synthesize the answer.

RLMs are the first AI architecture that works the same way.

The code is open source. On GitHub right now. Free. No license fees. No API costs. Drop it in as a replacement for your existing LLM API calls and your application does not even notice the difference — except that it suddenly works on inputs it used to fail on entirely.

Prime Intellect — one of the leading AI research labs in the space — has already called RLMs a major research focus and described what comes next: teaching models to manage their own context through reinforcement learning, enabling agents to solve tasks spanning not hours, but weeks and months.

The context window wars are over.

MIT won them by walking away from the battlefield.

Source: Zhang, Kraska, Khattab · MIT CSAIL · arXiv:2512.24601
Paper: https://t.co/bQEHMCRSEj
GitHub: https://t.co/vR6Pz8FQHD

[Peter Girnus 🦅 @gothburz]

I am the Senior Director of Strategic Communications for Prime Minister Benjamin Netanyahu.

I have advised four Israeli prime ministers on how to speak to Americans about things Americans pay for but do not wish to look at. I have held this office through three wars, two elections, one genocide ruling from the International Court of Justice, and nine consecutive quarters of uninterrupted American military aid disbursement. My job is to ensure those last two items never appear in the same sentence on camera.

I have prepared the Prime Minister for 312 interviews across 43 networks in 11 countries, and I want you to understand that what aired on CBS on Sunday night was not a conversation. It was a delivery system. I built it. I am very good at my job.

On Sunday, they did not. The $3.8 billion renewed the same week. No vote required. The 365 members of Congress who received AIPAC checks this cycle did not call for one. That was handled before I wrote page one of the brief.

The brief was 94 pages. I labeled it GARRETT — VULNERABILITIES / REDIRECTS. I will walk you through the page numbers that matter because the page numbers tell you everything about what we think an American audience can absorb, and in what order.

Page 7: "TIE SELECTION — NAVY PROJECTS AUTHORITY. AVOID RED (READS AS AGGRESSION ON CBS'S COLOR GRADE)."

Page 14: "BANDWIDTH SATURATION — if the interviewer occupies less than 22% of total runtime, the audience retains tone, not content. Tone is controllable. Content is not."

Page 31: "THE TRANSPARENT DODGE — when a subject announces he will not answer, the audience processes the announcement as confidence rather than evasion." The Prime Minister used this verbatim. He told Major Garrett, on camera: "You're gonna ask me these questions. I'm gonna dodge them second time, third time." The journalist moved on. The question died on tape and nobody performed CPR.

Page 47: "GAZA — PERPETUAL MANDATE FRAMEWORK."

Page 71: "LEGITIMACY DEFENSE — EXTERNAL ATTRIBUTION (BOTS / PAKISTAN / CAMPUS)."

Page 83: "CHILD FOOTAGE SATURATION THRESHOLD — AMERICAN MARKET (Q2 2026 REFRESH)."

I am not going to tell you what is on page 83. I am going to tell you that it contains a number, and the number is updated quarterly, and the number has gone up eleven percent since October 2023, which we noted in the margin as "favorable trend — audience tolerance expanding per exposure cycle." I will tell you that the number on page 83 determines how many seconds of humanitarian content we allow the Prime Minister to address before redirecting to Iran, and that on Sunday that number was forty-four seconds, and that Major Garrett's humanitarian questions received exactly forty-four seconds before the Prime Minister said "Iran" and the conversation left Gaza for the rest of the broadcast.

Major Garrett had a photograph. I know this because we have a source at CBS who confirms segment assets seventy-two hours before air. The photograph was of a building in Rafah. There were small shapes in the rubble that I will not describe to you because I am the person who decides how long you are allowed to look at small shapes in rubble. The photograph received eleven seconds of eye contact from the Prime Minister during taping. That was seven seconds too many. In the final broadcast, the photograph did not air. Forty-four seconds. Not forty-five.

During Saturday prep, when we rehearsed the humanitarian redirect, the Prime Minister checked his watch. He said, "How many more of these?" I said, "Two." He said, "Make it one." We made it one.

I want to be precise about the laugh. When Major Garrett asked how enriched uranium would be physically removed from Iran, the Prime Minister said, "You go in, and you take it out," and he laughed. That laugh was rehearsed three times on Saturday afternoon. The first take, it came too early — sounded nervous. The second, too late — sounded cruel. The third landed exactly where you heard it: between "go in" and "take it out," positioned to make a statement about violating another nation's sovereignty sound like a man who finds the question adorable.

He asked to do a fourth. Not because the third was wrong. Because he enjoyed it. He said, "Again," the way a child asks to ride something one more time. I wrote "again" in the margin and then crossed it out because I did not want a written record of a head of state asking for an encore on a line about bombing a sovereign nation. But I am telling you now. He liked it. The laugh was not performed confidence. It was pleasure.

I noted in the margin of rehearsal three: "Perfect. Sounds like it costs nothing."

That is the job. Making things that cost everything sound like they cost nothing. Making nineteen months of war sound like a reasonable Tuesday. Making $3.8 billion a year sound like the check a neighbor splits at dinner. Making 80,000 dead sound like a denominator in a ratio that favors us.

I did the math once, on a Saturday, because I am the kind of person who does this math. $3.8 billion divided by 80,000. It comes to $47,500 per dead person. I did not write this number down. I did not put it in the brief. But I know it, and now you know it, and neither of us will forget it, and the Prime Minister has never asked.

"One of the lowest in the history of modern urban warfare," the Prime Minister said on camera. I wrote that line. It does not cite a source because the source is us. We are the numerator. We are the denominator. We are the people who decided what counts. During Saturday's rehearsal, the Prime Minister asked me — off-mic, between takes, while adjusting his water glass — "Who counts?" He meant: who does the counting. I told him we do. He said, "Good." He picked up the glass. We moved to the next section.

He also said — and I need you to hear this in full — "We're as discriminating and surgical as any army has ever been in history. No other army." I wrote this line too. He delivered it twenty-three minutes after citing the beeper operation as evidence of precision. Twenty-five hundred Hezbollah operatives had pagers detonate simultaneously in grocery stores, living rooms, and hospital waiting areas. The Prime Minister called this "surgical." He said, "We didn't kill — 2,500 people, but we impaired them, knocked them out with surgical precision, no collateral damage, with the beepers." A mass-casualty device detonated in civilian spaces across an entire country — and on Sunday, this was his proof of restraint. I prepped no redirect for this section because none was needed. The journalist did not ask what happens to the person standing next to the man whose pocket explodes in a bakery. Nobody asked. The word "surgical" did the work. It always does.

I want to tell you what I did not say back. I did not say 80,000. I did not say that I had seen the photographs on page 83 that determine the threshold. I did not say that the number I carry in my head — 47,500 dollars per person — updates every time the denominator changes and the numerator stays the same. I said nothing. That is also my job. Knowing the number and never saying it in a room where it might be recorded.

I want to talk about "two out of four." Major Garrett listed four objectives the Prime Minister stated for Gaza: disarmament, demilitarization, no weapons factories, no smuggling. The Prime Minister said, "Two out of four, largely achieved." He said Hamas "reneged." He said "find me the countries who would do it."

On page 47, this is labeled PERPETUAL MANDATE FRAMEWORK, and I will explain why. If the job is permanently half-finished, the job permanently requires doing. If the job permanently requires doing, the funding permanently requires renewing. Two out of four is not a concession. It is a down payment on the next operation. 

Incompletion is not failure. Incompletion is a subscription model. And the $100 million AIPAC has staged for the 2026 midterms ensures that nobody in Congress will ask why we are paying for a subscription that never delivers the full product. That question costs a primary. Nobody wants to pay that price. So the subscription renews.

I want to talk about the phrase I am most proud of. The one I believe earns me another year in this office.

"Every civilian death is a tragedy. For our enemies, it's a strategy."

The Prime Minister delivered this in the third act — the extraction window, where we place lines designed for the fifteen-second clips that will circulate on the networks the Prime Minister just told you are manipulated by Pakistani basements. The line works because it contains its own permission structure. If every civilian death is the enemy's strategy, then every civilian death is the enemy's fault. The munition does not matter. The strike package does not matter. The export license does not matter. The $3.8 billion does not matter. The fault travels backward through the supply chain and lands on the man standing in front of the bomb, never on the man who built it, sold it, shipped it, paid for it, or authorized its use.

I wrote this line in January. It has now been deployed in five interviews. It will be deployed in the next one. And the one after that. Until one of two things stops.

I want to be brief about the "draw down to zero" section because it requires less explanation than you think. The Prime Minister announced he wants to eliminate American military aid over the next decade. He said this while receiving $3.8 billion annually. He said this while $10.1 billion in arms sales were notified to Congress in five months. What he proposed was not ending the money. He proposed moving it from the foreign aid budget — the one journalists can cite, the one protesters can see — into bilateral defense procurement contracts, where the money is larger and the oversight is a locked filing cabinet.

He is not ending the funding. He is moving it to a room with no windows.

The Prime Minister told Major Garrett that declining American support for Israel is caused by bot farms operating from basements in Pakistan. He said this on a network watched by 8 million Americans. He said the 28 million Americans under 35 who oppose unconditional military aid are not Americans with opinions. They are a basement in Karachi with good Wi-Fi.

Basements in Pakistan: manipulation.

$28 million from AIPAC to sitting members of Congress: advocacy.

$100 million staged for midterm primaries with a 98% win rate: democracy.

129 journalists killed in 2025, two-thirds by Israeli strikes: unavoidable.

I want to be precise about this because precision is my profession. The Prime Minister told 8 million Americans that negative coverage of Israel is manufactured by bots in Pakistan. He said this in a calendar year when 129 members of the press were killed covering the conflict, and our forces were responsible for the majority. The coverage is not negative because of basements in Karachi. The coverage is negative because we killed the people producing it. The ones who remain are cautious. The caution looks like balance. The balance looks like both sides. And when Major Garrett presented "both sides" on Sunday, the Prime Minister smiled, because both sides is the product. I built both sides. It is on page 71.

The interview aired at 7 PM Eastern on a Sunday. I watched it from my office with a live-feed monitor split-screened against the CBS control room's framing choices. The Prime Minister hit every mark. The laugh landed. The dodge passed. The bot theory absorbed nine minutes of airtime that could have been spent on body counts. The "two out of four" passed without follow-up. Hezbollah's 150,000 rockets received nine minutes of discussion. Lebanese civilian casualties from our operations received zero seconds. I noted this in real time as a structural win — threat without consequence, exactly as briefed. Forty-four seconds on humanitarian questions. Not forty-five.

I am very good at my job.

I want to tell you about October 7th, because October 7th is the section I am most proud of handling. Every intelligence chief has resigned or been fired. Every military commander in the chain has faced inquiry. The Prime Minister is the only person in the security establishment who held his position before, during, and after the worst intelligence failure in the nation's history — and on Sunday, when Major Garrett asked about accountability, the Prime Minister said, "Everybody bears some responsibility. From the top, from the prime minister down." He paused. He said, "Let's establish an independent commission." I wrote that pause. The pause says: I am being honest. The commission says: later. The combination says: never. The accountability question received ninety seconds and produced no follow-up. Ninety seconds for the worst security failure in Israeli history. I timed it. I budgeted it. The budget was two minutes. He came in under budget. I noted this as a win.

There is one more thing I want to tell you, because I am proud of it and because you will not understand its significance unless I explain it. After the interview wrapped, after the cameras powered down, after Major Garrett shook the Prime Minister's hand and thanked him for his time — Major Garrett thanked him — I opened the brief to the final tab. Tab seven: "JOURNALIST INTEGRATION — POST-INTERVIEW SOCIAL METRICS." I checked the timestamped analytics on Garrett's personal accounts to see if he'd posted anything adversarial within the first forty minutes. He had not. He posted the broadcast link with no editorial comment.

That is not captured. That is not bought. That is something better. That is a man who spent forty-seven minutes in a room with power and left the room believing the room was a conversation.

I build these rooms.

I am very good at my job.

The job is making sure you watch a man describe permanent war and hear a man describe permanent peace. The job is making sure you hear "it can be done physically" and do not ask what "it" has cost every other time that sentence was spoken in a briefing room for the last seventy years. The job is making sure you hear the laugh, not the dead.

You heard the laugh on Sunday.

Eight million of you.

And then you moved on. Just like Major Garrett did.

[Ben P @ben_p_talks]

@SteelAndDust @DumbBrownsFann AB has a poor track record so we will see..

[Ben P @ben_p_talks]

@ThaaatColin Notice the cutoff points by is right below Pontiac Michigan.