Therapist

I stepped into Web3 in April 2025, and within three weeks, I landed a job with no connections, no referrals, just belief, effort, and the courage to show up. Fast-forward to today, I can’t even come on this platform freely to post, engage, and grow my presence the way I should.

I just claimed my Early Access spot on @claudersfun. Limited spots — get yours: clauders.fun/?invite=2MJ3H x.com/claudersfun/st…

@aakashgupta Accurate atoms claim, but overstated. Body doesn’t fully renew. DMN changes self-experience, not self-deletion. Philosophy is interpretive, not empirical. Self is a brain-generated model, not an entity that disappears.

The meme is closer to physics than philosophy. Every carbon atom in your body was forged in the core of a star that died billions of years ago. Every atom heavier than helium. Your bones, your blood, your DNA. All of it cooked inside stellar fusion, blown across the galaxy in a supernova, and accreted into a planet before it could become you. There are roughly 7 octillion atoms in your body. Not one of them was made on Earth. Now run the boundary. You swap trillions of atoms with the air around you with every breath. Your skin fully replaces itself every 27 days. Most of the atoms in you right now weren't part of you a decade ago. There is no atom you can point to and say "this has always been me." The body is a standing wave. Stuff is constantly flowing through. The brain part is wilder. The sense of being a separate self gets generated by a specific network called the default mode network. Suppress it with psilocybin, deep meditation, or general anesthesia and the felt boundary between self and world collapses entirely. fMRI scans of people in ego-dissolution states show reduced activity in exactly the region that was simulating "individual." The brain manufactures the user. Turn off the manufacturing process and the user disappears. Vedanta worked this out around 800 BC. Spinoza wrote it down in 1677. Schopenhauer called the ego a trick of perception. They were running on intuition. Physics and neuroscience just put numbers on it. Stardust, pretending.

@Mayorwhistle10 Bare times Man, I'm sure there's no way we' going back to that

Yes during my 100 lvl first semester..😂

@Mayorwhistle10 Wtf 😂... It's gone?

Sorry bro, it's gone. 💔

The transformation from Nobody to Somebody starts first with a mindset shift . But from Somebody to a Nobody only needs a shift in focus.

@aakashgupta The real shift is from image capture to human guidance. Once hardware matured, the bottleneck became behavior. Huawei is betting pose intelligence becomes as standard as HDR, and that looks plausible given it has already teased AI pose guidance for the Pura 90.

Smartphone cameras went from 2 megapixels to 200 megapixels over 15 years. Posed photos of actual people barely got better. The camera was never the bottleneck. Every professional photoshoot runs the same way. A director behind the lens tells the subject where to put their hands, where to angle the chin, how to shift weight onto one leg. Photographer and subject work as a pair. Remove the director and the best camera on Earth still produces stiff, awkward photos of stiff, awkward people. Huawei just shipped that director in software. The Pura 90 launches April 20 with a feature called AI Posture Recommendation. You open the camera, tap one button, and a white outline appears in the viewfinder showing the subject exactly how to stand, sit, and hold their hands. Align yourself to the outline. Take the shot. Google shipped Camera Coach on the Pixel 10 eight months ago. It teaches the person holding the phone. Huawei went the opposite direction. They're teaching the person being photographed. Selfies and group photos dominate what gets shared online. The subject is the higher-leverage intervention and the bigger market. The V-sign is the clearest proof of the gap. Huawei's own promo mocks it. An entire generation of East Asian kids learned to throw up two fingers in every photo because nobody taught them anything else to do with their hands. The peace sign has been the universal fallback pose for 40 years. Software was always going to replace it. Every decade, a new camera capability stops being a feature and becomes the default. HDR. Portrait mode. Night mode. Posing is next.

@rubenhassid The subscription fee is the clause in my Claude setup. I'm still using the free tier, sadly

@rubenhassid This is the shift most people still miss: Claude is no longer just a chatbot, it’s becoming an operating layer for work. The real edge is not better prompting. It’s turning repeat tasks, context, and execution into a system.

You're using Claude like it's 2024. Here's the full 2026 setup (one cheatsheet to save): 1. Get the Claude app (claude .com/download). 2. You need a Pro plan ($20/mo). Worth every cent. 3. You'll see 4 tabs: Chat, Cowork, Code & Projects. ----- ✦ Cowork = Claude works inside your computer. Pick a folder. Add 3 files: about-me .md, anti-ai-style .md, my-company .md. Go to how-to-ai.guide to my .md files. Don't pay anything, get it in the welcome email. Set Global Instructions so it reads before tasks. Claude generates clickable forms. Click & answer. ✦ Projects = persistent context for your team. Go to claude .ai → Click "Projects" in the sidebar. Upload your brand docs, writing samples & data. Set custom instructions. Chat inside the Project. Every conversation remembers everything. ✦ Skills = slash commands that fire automatically. Open Cowork → prompt: "Use the skill-creator to build a skill for [your task]." Claude interviews you. Builds the Skill. You upload it once. Now type /brief anywhere. It just works. Browse plugins: Customize → Personal plugins → +. ✦ Code = a junior developer for $0.66/day. Go to the Code. Connect your free GitHub account. Type what you want in English. Attach a screenshot. Claude builds it. Pushes it live. You edit with words. Install VS Code + bypass permissions to go 10x fast. ----- Pro tip: Select Opus 4.6 + Extended thinking for complex tasks. Use Sonnet for quick ones. Help others be better at Claude ♻️ repost this. PS: I write a newsletter to 484,000+ readers. ↳ Join us for free at how-to-ai.guide. ↳ Open the welcome email after subscring. ↳ Download all of my infographics, also for free.

@aakashgupta Sports rights are no longer media products, they are paid acquisition channels. The antitrust angle matters, but economically the leagues already chose the highest bidder. If universal access is the deal, Congress has to force it, because the market clearly will not.

Peacock paid $110 million for ONE NFL wild card playoff game in January 2024. That single game tripled Peacock's quarterly subscriber growth. 3 million net-new accounts in Q1 2024, up from 1 million in the same quarter the year before. Amazon pays $1.5 billion a year for Thursday Night Football. YouTube pays $2 billion a year for Sunday Ticket. Netflix pays around $150 million for two Christmas Day games. The NFL's current 11-year media package projects over $100 billion in total rights fees, the largest sports deal in U.S. history. That's 2-3x what linear TV was paying per game. The leagues aren't making fan-hostile decisions by accident. They're running an auction and streaming keeps outbidding cable because every game is a subscriber acquisition machine. Peacock's $110M drove roughly 2 million incremental subscribers in a single quarter. At $7.99 a month, the retention math breaks even before the sub's first anniversary. The sports bar was never in that auction. Bars pay Comcast or DirecTV for commercial cable packages at inflated rates, and a slice of that flows up the chain to the league. Streaming services don't have commercial licenses. Amazon, Peacock, YouTube built their entire pricing stack around individual consumer subscriptions, because that's where the subscriber-growth metric Wall Street values them on lives. The leagues don't miss the bar revenue because most of it was never coming to them directly. Cable was the middleman, and the middleman is what streaming disintermediated. Senator Mike Lee wrote to the DOJ in early March arguing the 1961 Sports Broadcasting Act was passed on the assumption games would be financed through advertising and made available free to the public. The DOJ opened its investigation this month. The argument is that paywalled sports break the antitrust exemption Congress gave the leagues in exchange for universal access. The public shared-viewing experience was an accident of 1960s broadcast law, extended for 40 years by cable's commercial licensing tier. Both layers are collapsing at once. The NFL is about to collect $100 billion in rights fees, and there's no commercial Peacock account the bar down the street can legally buy.

@SamuelXeus @meekdonald Being right on direction means nothing if the trade structure is wrong. He nailed the top, but 100x leverage + deeply negative funding turned the position into a time bomb. In perp markets, entry is only half the trade. Carry cost can kill you before price proves you right.

He shorted the exact top of $RAVE but still ended in a loss, how does this happen? @meekdonald, a student I am actually proud of made a mistake I think a lot of people make and paid a massive price for it Although it’s sad, I have decided to use this avenue to help you all never fall for this error again First, what are funding rates? Funding rate is the price paid every few hours by traders to other traders to prevent the perpetual futures price from drifting away from the token spot price Think of it like a toll. If too many people are driving in one direction, they pay a fee. That fee goes directly to the people driving the other way So what exactly happened with $RAVE? $RAVE is the token for RaveDAO that was trading at 0.25 earlier this month, then the price ran to an all-time high of $19.54. That is a 6,000% gain in one week. Everyone will definitely start paying attention at this point The issue came when RAVE didn’t stop pumping, which meant the token accumulated so many shorts at the ATH that the funding rate flipped negative - Positive funding rate meant you get paid if you are on a short - Negative funding rate meant you get paid if you are on a long If you have being following, you’d realize the fee comes from the opposite team For Positive funding rate, the long gets to pay the shorts For Negative funding rates, the shorts gets to pay the longs @meekdonald who was lucky enough to short very close to the ATH, did so when the funding rate was at -2% I’d explain 👇🏾 With a leverage of 100x, an $100 position is valued at $10,000 and with a -2% funding rate, he is to pay $200 every 8 hours which is more than his margin size With his first funding rate payment, he is already in a 200% loss which obviously continues. So what’s the way forward? Before you enter a token that is running wild, make sure you check the funding rate Most exchanges show it directly on the trading screen. Here is where to look: On the futures trading page: • Binance: top of the chart near the price, shows current rate and a countdown timer to next payment • Bybit: same, right next to the mark price • Hyperliquid: visible on the trading interface near your position • OKX: shown in the contract details section It usually looks something like this: Funding Rate: -1.37% | Next funding: 04:32:17 That countdown tells you exactly how long until the next payment gets taken or received. Third party sites if you want to compare across exchanges: • Coinglass - shows funding rates for the same token across all exchanges side by side The key things to check before entering a trade: 1) What is the current rate 2) How long until the next payment 3) Is it positive or negative 4) How long do you plan to hold versus how many payments that means If the rate was extreme like RAVE was, even checking it once before entering would have told you this trade has a very short survival window Now you know. Tell me thank you 😁🫵🏾

@_ayotomiwaa @yurcoach @t3_kit @TheSultaan001 @Godonlyson2 @Cyb3erSage The starting point to wealth generation is knowledge accumulation... Big ups Guurl

Day 1 of learning Growth strategy with @t3_kit Does Web3 growth really come from numbers or community? 1/3 I used to think growth was all about getting more users or followers But Web3 is teaching me something different

@aakashgupta Apollo didn’t win by making the database smarter. They won by wrapping a low-retention commodity inside a high-frequency workflow. That’s the real AI moat in SaaS: not better output, better habit formation. Retention is where valuation multiples actually get defended.

Apollo.io's users would download 21,000 contacts and never log in again. One of the best B2B contact databases in the world, and the dominant usage pattern was single-session. Download. Cancel. The product was a commodity with SaaS pricing and zero retention characteristics. The AI strategy that took Apollo from $750M to $2.5B valuation was built on converting that single-session commodity into a daily workflow tool. Three AI products, sequenced. AI email writer to generate outbound. AI email warmup to build deliverability for new email addresses before they ever sent a real message. AI email responses, still in progress. The warmup feature failed the first time. The AI-generated emails were too absurd. They shelved it completely. Months later, as foundation models improved, they relaunched with better fine-tuning and RAG pipelines. The relaunched version worked. 10-20% retention improvement across the engagement product suite. The $1.75B in valuation gain traces back to one strategic insight: the contact database was the acquisition engine but AI-powered email tools were the retention engine. And in SaaS math, retention compounds. Acquisition doesn't.

@aakashgupta Exactly. Most “knowledge management” failed at cognition-adjacent labor, not cognition itself. Humans are good at insight, bad at relentless document hygiene. LLMs flip that equation. The real unlock is turning stale docs into living memory systems.

Every knowledge base ever built for humans died for the same reason. The maintenance job went to the wrong person. Cross-referencing sources, updating summaries when new data arrives, flagging when a new finding contradicts an old one. That's a specific, tedious, never-ending job. Nobody does it. So the knowledge base rots and everyone goes back to starting from scratch. Karpathy's fix: give that job to the LLM. You dump raw sources into a folder. The AI compiles a wiki. Writes summaries, builds cross-references, flags contradictions, maintains an index. You never touch the wiki. The AI maintains it entirely. 18.7M views on his post. The insight underneath is simple: the problem was never the reading or the thinking. It was the bookkeeping. LLMs don't quit bookkeeping.

There is an art to collapsing slowly. To making devastation look deliberate. To folding under the weight with such grace that people mistake it for dancing. I have mastered that art. I can carry a graveyard in my ribs and still sound composed. This Art can't!

@defileo The real shift is not “7 Mac Minis.” It is turning freelance labor into a parallelized system. Most people are still selling hours. He bought throughput, consistency, and time back. That is when a service business starts behaving like software.

He bought 7 Mac Minis for $4,900, his roommates thought he lost his mind. 9 months later he lost his mind at how much money he was making. He was doing freelance outreach manually every single day, scraping leads, writing emails, following up, updating spreadsheets at midnight. $4,500 a month, 60 hours a week, $18 an hour, less than the guy making coffee downstairs. So he bought 7 Mac Minis at $700 each. $4,900 once, that's it. His roommate walked in and asked what he was doing, he said "going to sleep" and he did. While he slept the machines worked. One found leads, one wrote personalized emails, one followed up, one updated the CRM, one handled scheduling, one monitored everything overnight, one reported the results every morning. He woke up to 600+ processed leads and 14 replies waiting, hadn't touched his laptop since the night before.

@aakashgupta This is what an AI talent war looks like when money stops being a constraint. Startups can raise billions and still lose if Big Tech is willing to price key researchers like sovereign assets.

Meta tried to buy Thinking Machines Lab for ~$1 billion. Murati said no. So Zuckerberg did the next best thing. He bought the founding team, one person at a time. This is the fifth founding member to leave for Meta in 14 months. The first, Andrew Tulloch, reportedly got a package worth up to $1.5 billion over six years. $250 million per year. For one engineer. Three more founding members went back to OpenAI, including CTO Barret Zoph. That's 8 founding team members gone from a startup that launched with about 30 people and raised $2 billion at a $12 billion valuation. Murati replaced Zoph with Soumith Chintala, the creator of PyTorch, and quietly hired Neal Wu, one of the best competitive programmers alive. The company also signed a deal with NVIDIA for one gigawatt of Vera Rubin compute. She's rebuilding the plane mid-flight. But the pattern reveals something about AI in 2026: the floor for retaining elite talent is nine figures per person, per year. $2 billion in funding sounds like a lot until you realize Meta will spend more poaching five engineers than most AI startups raise in their entire lifetime.

@quionie Nike is selling a lifestyle; the persona of having uninhibited ability to do anything just by thinking about it. So yeah, if they introduce an AI agent tomorrow, they'd just make it feel more super doper than most. And people will always buy into the possession of matchless power

@eliana_jordan Since the advent of AI agents and Automation, X has turned almost as professional-toned as Substack. I'm hoping that in the nearest future we no longer need LinkedIn for anything but only X

Had a conversation with a friend about AI and apps yesterday. She said: “Wow… how do you know all of this?” On X, I always feel behind. Like everyone knows more. Builds faster. Markets better. Like there’s no room left. Then you step outside the bubble… Talk to normal people… And realize: You’re not behind. You’re actually way ahead. Keep going. You’re doing better than you think.

@aakashgupta If those numbers hold, the story is not “AI is expensive.” It is that enterprise AI with real workflow embed can compound faster than consumer hype. Anthropic is being valued less like a model lab and more like a new operating layer.

Anthropic just turned down $800 billion. Two months ago they raised $30 billion at a $380 billion valuation. Now VCs are lining up to invest at more than double that, and the company is resisting. The math tells you why. Anthropic hit $30 billion in annualized revenue last week. Up from $9 billion at the end of 2025. In four months, they added more revenue than most Fortune 500 companies generate in a year. $1 billion ARR in January 2025, $30 billion in April 2026. No company in American history has scaled organic revenue this fast at this level. At $800 billion, investors are pricing Anthropic at roughly 27x current revenue. Sounds rich until you plot the trajectory. Brad Gerstner of Altimeter said on All-In last week that Anthropic could exit 2026 at $80 to $100 billion in revenue. If that's even close, $800 billion at year-end revenue is 8 to 10x. That's cheaper than Salesforce. The enterprise numbers explain the acceleration. Over 1,000 businesses are each spending more than $1 million annually on Claude. That number doubled in less than two months after the Series G. Claude Code alone is running $2.5 billion in annualized revenue, and enterprise users make up more than half of that. Anthropic surpassed OpenAI in revenue last week. $30 billion vs $25 billion. The company with fewer consumer users and no ChatGPT-level brand awareness is generating more money. 80% of Anthropic's revenue comes from enterprise. Higher retention, lower churn, contracts that expand over time. The VCs offering $800 billion think they're paying a premium. Anthropic knows they'd be selling a discount. That's why they said no. The IPO is where this reprices.

I built 7 Claude Skills that replaced 6+ hours of weekly marketing work. Now I’m giving them all away. The Founder Skills Vault: → Viral Content Generator → Content Pillar Generator → LinkedIn Content Analyzer → Lead Magnet Idea Generator → LinkedIn-to-X Converter → Warm DM Strategist → Lead Qualifier This isn’t a prompt pack. These are production-grade tools that run inside Claude. It’s the same system I used to grow 7,000+ followers in 3 months. Want it? → Like + RT → Comment “FOUNDER”