Corporations are starting to ban AI for employees ❌

While people fear AI will take their jobs, big companies are facing the opposite problem: after mass adoption, they are starting to limit AI usage internally.

💻 Microsoft is reportedly canceling most Claude Code licenses for engineers and moving them back to GitHub Copilot CLI. Why? Token costs got too high

💻 Nvidia sees the same issue. VP Bryan Catanzaro said compute costs for his team already exceed the cost of the employees themselves

🚖 Uber is the craziest case: the company burned through its entire 2026 AI budget in just 4 months. 84% of its 5,000 engineers used Claude, 70% of new code was AI-generated, and top users were spending $500–$2,000/month each. Even Uber’s CTO spent $1,200 in one two-hour demo

So the tool that was supposed to save money became expensive at scale

The next AI race may not be about who has the smartest model, but who can make AI cheaper to use. Winners won’t be those with the most advanced AI – they’ll be those who use it efficiently and economically 🧠

Is the AI Rally a Bubble or a Boom 🤔

The chart captures the core distinction between what ends badly and what usually does not. In a bubble, price runs far ahead of earnings until the gap closes. In a boom, earnings grow fast enough to carry the price.

Lately, there has been growing talk that the stock market is overheated, that the growth driven by AI has run its course, and that tech companies are significantly overvalued. Let’s take a critical look at the situation.

🔽 Back in 2000 Cisco peaked near 200x earnings, and Pets. com went public with none at all. The entire case for internet stocks rested on revenue from a technology that was years from generating the cash flows the market had already priced in. When that gap closed, the Nasdaq fell 78% and took 15 years to recover.

🔍 Today's core companies look different. Nvidia's revenue went from $27B in fiscal 2023 to $130.5B in fiscal 2025, up 114% in a single year, with net income reaching $74B. The stock has run roughly 2,000% since its 2022 lows, but earnings grew fast enough to keep the forward P/E near 30x. The Nasdaq-100 forward P/E sits near 26x today versus roughly 50-60x at the 2000 peak.

Companies are spending hundreds of billions on GPU and data center capacity before enterprise use cases have proven their economics at scale. The earnings are real, but they depend on that spending continuing, and that spending needs to be justified by what the AI actually produces for businesses paying for it.

👁️ Nvidia's earnings track directly with infrastructure spending. If enterprise buyers start questioning the ROI on their GPU deployments, the spending slows and the earnings story changes with it.

But so far, every quarter that was supposed to show the slowdown has shown the opposite. Until the earnings stop growing, this looks more like the right side of the boom chart than the left side of the bubble one. Just my opinion...

#FAQ

The best Hyperliquid monitoring stack 🔍

The broader market is bleeding while $HYPE just printed a new ATH today. The token of the most successful DEX in crypto has outperformed Bitcoin for 2 years running now.

Been getting DMs asking what I actually use to track traders on Hyperliquid and other related stuff. So I decided to make a small list of cool tools I use 👇

🟢HL.eco: full picture on HYPE tokenomics: buybacks, burns, fees, and protocol revenue.

🟢Hypurrintel: tracks HYPE spot ETF flows. Useful for watching where the bigger money is moving.

🟢Hyperdash: cohort analysis of traders ranked by profitability, from the worst to the most consistently profitable. Also works as a trader tracker: open positions, trade history, full breakdown.

🟢Hypurrscan: blockchain explorer for Hyperliquid. Covers on-chain stats, HyperEVM, and HYPE staking.

🟢Dune: full-stack dashboard covering market structure, trading volume, open interest, user growth, fee generation, and validator economics. Best single view of Hyperliquid as a live financial system.

If you haven't heard about Hyperliquid yet or don't know how to use it, read this post 🫲

@Coin_Post

The case against all of DeFi 😨

Manuel Aráoz, founder of OpenZeppelin, posted a short and scary message: all DeFi is now unsafe. His reasoning is that AI coding agents are superhuman at finding vulnerabilities, and the math is brutal. Defenders have to fix every bug while attackers only need one.

Aráoz helped build the security infrastructure most of DeFi runs on before leaving as CTO in 2019. He understands this code at a level most people don't, and the asymmetry argument deserves to be taken seriously 👀

But let’s try to assess the situation calmly and not give in to panic. Look at the actual exploit record: the vast majority of major DeFi hacks weren't zero-days buried in audited code. They were compromised keys, misconfigured access controls, and privileged role exploits.

🤔 Even OpenZeppelin itself pushed back on Aráoz today, saying most recent incidents traced to operational failures, not smart contract bugs.

We know for sure that, for example, Aave's core contracts have been reviewed by thousands of developers over years, and it's very simple. A sleeper exploit sitting in that code is just not possible.

😱 Where he's probably right is everywhere else. New protocols, smaller apps, anything that hasn't been battle-tested for years, the asymmetry is real. And it cuts deeper because of how interconnected DeFi is. One weak link in a protocol that plugs into everything else is enough to spread the damage far beyond the original exploit.

The practical result is that DeFi gets less creative. If deploying novel code means handing an AI-powered attacker a fresh target, most teams will stop trying. What survives for sure will be forks of apps like Aave and Uniswap, which is safer, but the experimental edge that made this space interesting goes with it 🫡

@Coin_Post

Trump wants his face on a $250 bill 💵

The Trump administration is pushing the Bureau of Engraving and Printing to design a $250 note featuring Trump's portrait. If it launches, he'd be the first living person on US currency since 1866, when the practice was banned after a mid-level Treasury bureaucrat appeared on a 5-cent note.

🤔 There's a legal obstacle: federal law only allows deceased individuals on bills, and a $250 denomination isn't authorized without an act of Congress. The bureau's director kept raising these objections. She was abruptly reassigned last month. Her replacement is one of the political appointees who was pushing for the note.

Most people are fine seeing their face in a mirror once a day. Some need a little more 🤦

The best free transcription AI tool most people have never heard of 🤖

OpenAI open-sourced Whisper in 2022 and most people still don't know it exists. It transcribes audio and video locally on your computer, with near-human accuracy, no subscription, and no file size limits.

🤖 The easiest way to get started: on Mac, download MacWhisper. On Windows, download Whisper Desktop. Both are free GUI apps built on top of Whisper, drag in a file, pick a model, get a transcript. No terminal, no setup. The base model runs on any machine. The larger models are more accurate and still free.

The privacy angle is the same as with Ollama. Most transcription services upload your files to their servers. Whisper runs entirely on your machine. For interviews, client calls, or anything sensitive, that matters. I've used it on hour-long recordings and the output needs maybe 2-3 small corrections 🎧

If you're paying $10-15 a month for a transcription service and not doing anything that specifically needs live meeting notes or speaker diarization, this covers the same ground for free.

💸 There are plenty of AI services that are free or nearly free, and there are tons of opportunities out there. If you're interested, be sure to check other posts with the hashtag #AI

Bonk is running a waitlist for something unannounced, with a drop confirmed for day ones 🪙

The whole thing takes under a minute. Go to waitlist website and join with a Solana address, X account, or email. With three sign-up options, it's easy to create multiple accounts for waitlist, which is recommended 🧠

#earn

HYPE just hit a new ATH while everything else is still in the bear market 😲 ↗️

Bitcoin and Ethereum are down 40% and 60% from their highs. Crypto has been dead money for most of 2026. One token is doing the opposite. $HYPE hit a new ATH today of $67.5, and is up roughly 228% since January. With a $14.6B market cap, it's the 11th most valuable crypto asset in the world.

🤑 Hyperliquid Strategies, the largest HYPE treasury fund, is sitting on $1.2B in unrealized gains. Every single ETH treasury is deep in the red. BitMine, the largest ETH treasury company, spent roughly $16.4B accumulating its position and is now sitting on over $8B in unrealized losses. Their stock is down more than 63% in 6 months.

ETF flows are starting to favor HYPE too. Bitcoin ETFs saw over $1B in outflows last week and Ethereum funds lost another $215M, while HYPE spot ETFs drew about $72M.

🏀 If you've been reading this channel since late 2024, you already knew about HYPE. Hopefully some of you are sitting on nice profits right now

SpaceX IPO is designed to dump on passive investors 🤔

SpaceX is Elon Musk's rocket and satellite company, best known for Starlink and reusable rockets. The IPO also bundles in xAI, Musk's unprofitable artificial intelligence venture, which lost $2.47B in Q1 alone and accounted for 76% of SpaceX's total capital spending that quarter. Musk owns the whole thing and will retain 85.1% of voting power after the listing. The company is targeting a June 12 IPO at a $1.75T valuation.

🔍 But before that, index providers changed the rules. FTSE Russell now adds eligible large IPOs after just 5 trading days (was a full year). Nasdaq fast-tracks top-40 candidates after 15 trading days (was 3 months). S&P proposed cutting its seasoning window from 12 months to 6 and waiving profitability requirements that SpaceX would never have passed under the old rules...

Once SpaceX enters those indexes, every ETF and fund tracking them must buy, simply because it is in the index. At 100x trailing sales, all passive investors in retirement accounts will be buying one of the most expensive stocks in market history with no say in the matter. Even a Danish pension fund already blacklisted SpaceX over this. Index investors don't get that choice.

At $1.75T, the valuation prices in Mars missions and space AI data centers that don't exist yet. SpaceX lost $4.28B in Q1 and carries a $41.3B accumulated deficit. The rule changes guarantee mechanical buying from tens of millions of passive investors who never opted to own it. This is clearly a scheme designed to create the perfect conditions for insiders to cash out ❗️

@Coin_Post

What Is a Normal P/E Ratio 📊

P/E ratios vary by sector because growth profiles and margin structures differ. Banks and utilities typically trade at 10-12x. Consumer staples run 18-22x, tech companies with recurring revenue sit at 25-35x, and pure growth names go well above that when investors are buying future earnings rather than current ones.

❗️ Above 40-50x, a stock is either expensive or the market expects the business to look very different in a few years.

The chart shows six Magnificent Seven stocks names clustered between 23x and 38x with a an obvious 'imposter' next to them: Tesla is at 398.7x ➕

A P/E that high means current earnings are nearly irrelevant to how the stock is priced. Investors are not paying for what Tesla earns today as a car company. They are paying for a future version of the business: autonomous vehicles, robotaxis, humanoid robots...

🐂 Bulls argue those businesses will eventually generate earnings that make today's price look reasonable.

🐻 Bears argue the market has been pricing that in for years without it arriving. Both have been saying the same thing since 2021.

Above 50x, earnings stop being the right valuation tool and analysts switch to price-to-sales or free cash flow instead.

#FAQ

10 stocks are carrying the entire stock market 📊

Most people look at the S&P 500 and think markets are doing fine. What they're actually seeing is a cap-weighted index where Apple, Nvidia, and Microsoft dominate the calculation. When they go up, the index goes up, regardless of what the other 495 companies are doing.

🔍 The equal-weighted S&P 500 gives every stock the same slice. That version is up 34% since early 2023. The cap-weighted index is up 81% over the same period. The 47-point gap between them is the largest since the Dot-Com Bubble.

The ratio of equal-weighted to cap-weighted just fell below its 2008 Financial Crisis low, down 28% from its February 2023 peak. In 2008, the financial system was collapsing. Today the banks are fine. The market is just being carried by a handful of stocks, and almost everything else is going nowhere.

😳 When concentration gets this extreme, history offers two outcomes: Either the mega-caps drag the rest up, or they roll over and take the whole index with them.

@Coin_Post

Tokenization keeps going global 🟩

If Binance really enables conversion of US stocks into tokenized stocks on BNB Chain, this could be a big step for on-chain equities.

Non-US users would get easier access to US stock exposure directly through crypto wallets: fractional shares, 24/7 trading and faster settlement – without relying on a traditional broker.

Bigger picture: tokenized stocks are slowly turning from a niche narrative into real market infrastructure. And with Binance’s user base, liquidity and volumes in this sector could scale much faster 👍

Google Gemini + NotebookLM: how to build your own AI tutor 🧑‍🏫

Feed AI any study material, and it can turn it into a personal tutor that explains, tests you, finds weak spots, and helps you actually remember things.

1️⃣ Upload your materials to NotebookLM

▶️ It can be:
PDFs, YouTube videos, lecture notes, slide decks, articles, book chapters, research papers

2️⃣ Build a learning roadmap

▶️ Prompt:
Act as a world-class tutor. Turn these sources into a learning roadmap, not a summary.
Break it down into:
Big Picture: what this topic is about, why it matters, where beginners get confused.
Core Concepts: 5–10 key ideas, simple explanation, why each matters.
Learning Order: what to learn first, what comes next, what depends on what, what to ignore for now.
Difficulty Map: easy, medium, hard parts and what requires repetition.
Final Roadmap: clear step-by-step path with no fluff.
Goal: Make the topic organized and easy to study

3️⃣ Turn the material into a personal tutor

▶️ Prompt:
Act as my private tutor. Teach this material clearly.
For every major concept, explain:
Meaning in simple words.
Why it matters.
Real-life example.
Beginner-friendly analogy.
Common mistake.
Quick check question.
After each section ask: “Do you want a simpler explanation or should we continue?”
Tone: simple, patient, practical.
Goal: Help me understand, not just read

4️⃣ Create active recall questions

▶️ Prompt:
Create an active recall system from these sources.
Generate:
20 basic questions.
15 medium questions.
10 advanced questions.
5 trick questions.
5 explain-in-your-own-words questions.
5 real-world application questions.
For each question include:
• correct answer
• source concept
• what to review if I’m wrong
Important:
• do not show answers first
• ask in rounds
• wait for my reply
• grade strictly
Goal: Train memory through retrieval, not passive rereading

5️⃣ Find your weak points

▶️Prompt:
Here are my answers:
[paste your answers here]
Now grade me like a strict but helpful tutor.
Do NOT just give me the correct answers.
Instead, analyze my learning gaps deeply.
For every wrong or weak answer, explain:
1. What I got wrong
2. Why my answer is incorrect
3. What concept I misunderstood
4. The simplest correct explanation
5. The exact part I should review again
6. One new question that tests the same weakness
Then: Create a ‘Weak Points List’ ranked from most important to least important.
Goal: Show me exactly what I still don’t understand, and why, so I can fix it properly

6️⃣ Use Gemini to simplify the hard parts

▶️ Prompt:
Act as my learning coach.
These are the concepts I still don’t understand: [paste weak points]
Explain each concept in 4 levels:
Like I’m 10.
For a beginner.
For exam prep.
For real-life use.
For each concept include:
• analogy
• simple example
• common mistake
• mini exercise
• test question
Goal: Make hard concepts finally click

7️⃣ Build a 7-day study plan

▶️ Prompt:
I want to learn [topic] in 7 days.
My current level is: [beginner / intermediate / advanced]
My materials are: [paste what you uploaded to NotebookLM]
My goal is: [exam / work / project / general understanding]
Now build me a structured 7-day study plan using:
1. Active recall
2. Spaced repetition
3. Quizzes
4. Review blocks
5. Practical exercises
6. Final testing
For each day, include:
• What I should study
• What I should review
• What I should practice
• Questions I should answer
• Time needed for each section
• How I can check if I truly understood it
Rules:
• Keep it realistic and not overwhelming
• Focus on memory and understanding, not passive reading
• Include regular review of weak points
• Make it practical and easy to follow
Goal: Give me a clear, actionable study plan that I can actually stick to and complete successfully

Save this 📌

#FAQ

Trump says he'll never let crypto down 🇺🇸
Market makers: hold my beer 👌

BTC just lost $68k 🌡

@Coin_Post

Span, backed by Nvidia, is testing mini data centers that look like giant AC units and can be installed in regular homes 🤔

Homeowners get free high-speed internet and a fixed electricity bill of $150/month. Everything above that is covered by Span, saving them around $180/month.

For Span, paying someone’s utility bill is still way cheaper than building and running a full-scale data center.

The only sad part: the box has 16 Nvidia RTX 6000 GPUs inside, but you can’t mine crypto on it – Span’s control unit will make sure of that 😣

#AI

AI tools worth bookmarking 🧠

AI is way more than just ChatGPT these days. Here are a few tools that can actually save time across coding, design, content, video, and productivity:

▶️ Gamma
▶️ Creates presentations, documents, and landing pages from a simple prompt. Great for turning ideas into polished visuals fast

▶️ Cursor
▶️ AI-powered code editor that helps build websites, bots, scripts, and automations, even if you're not a developer

▶️ Runway
▶️ I video toolkit for generating clips, removing objects, replacing backgrounds, and animating images

▶️ Stitch
▶️ Google's AI tool for turning text prompts into app and website interfaces. Useful for rapid prototyping and product ideas

▶️ Magnific AI
▶️ Upscales images and adds realistic details, making ordinary visuals look much more polished

▶️ Recraft
▶️ Creates illustrations, icons, vector graphics, logos, and brand assets with a strong focus on design consistency

▶️ Napkin AI
▶️ Turns text into diagrams, visual maps, and simple infographics. Perfect for explaining ideas without opening Figma

📌 Save this and send it to someone who still thinks AI is just a chatbot

Farm a chance for 0.1 SOL with Umbra 💸

Umbra Privacy is a privacy protocol on Solana for confidential transactions. The project already has a token, wallet and active ecosystem.

To do:
🔵 Go to the website and join the waitlist
🔵 If it asks for a code manually – use the code after ref= in the link
🔵 After signing up, grab your own referral code/link and share it
🔵 Check your spot on the leaderboard
🔵 Watch their X – they may drop open access codes from time to time

They mention up to 0.1 SOL for getting access, but rewards will likely depend on leaderboard position and activity. A few accounts may increase the upside – just keep it clean and don’t spam 🤫
DYOR & NFA

#earn

Need a few more 🔥 on the post 👆

Hit 30 and I'll drop Part 2 with more tools that can save your wallet from expensive lessons

@Coin_Post

AI boom: trillions into the fire 🔥

Big Tech is deep in the red after spending around $1.4T on artificial intelligence: Amazon is down $291B, Google – $262B, Microsoft – $235B, Meta – $227B.

And somehow, with trillions already burned, the US tech sector still rallied 42% in the last 2 months – the strongest 2-month move in 24 years 💹

“Investing in the future” is getting expensive. Meanwhile, Nvidia has already made $253B selling everyone the shovels for the AI gold rush 🍪🍪