POV: claude traveled 6 months into the future and told you exactly how your next move failed.
it's called a premortem.
daniel kahneman (nobel prize-winning psychologist behind "thinking fast and slow") called it his single most valuable decision-making technique.
google, goldman sachs, and procter & gamble all use it before major launches.
here's the problem it solves.
when you ask claude "is this a good plan?" it finds all the reasons to say yes.
that's what it was trained to do. so you walk away feeling confident.
you execute, and spend weeks / months building on top of that plan.
then it blows up.
and you realize the problem was obvious in hindsight, you just never stress-tested it because claude told you it was solid.
a premortem fixes this by flipping the frame.
instead of asking "what could go wrong?" you tell claude "it's 6 months from now and this is already dead. tell me how it died."
that shift turns off claude's optimism because there's nothing to be optimistic about. the premise already says it failed.
so claude stops looking for reasons your plan will work and starts explaining how it fell apart.
claude comes back with every way your plan could die, each one with a full failure story and the early warning signs to watch for.
then a synthesis pulls it all together:
> which failure is most likely
> which failure is most dangerous
> the single biggest hidden assumption you're making (often the most valuable part)
> a revised version of your plan with the gaps closed
you say "premortem this" and give it your plan. the skill handles the rest.
How I have Claude automate anything I do on my computer just by showing it:
I use the free Clips app and hit record to start recording.
Then I just do the action - like navigating to Rippling and approving any unapproved time off - and Iโll often just talk out loud as I do it explaining any addโl details like I would when showing a human.
The Clips app captures everything on your screen and everything you say, and transforms it into a format Claude can โsee and hearโ.
You'll then get a link that you can easily share with Claude (or humans too).
Iโll paste it to Claude and say something like, "automate this for me every day at 8:00 AM."
The link has special metadata that agents like Claude can discover.
It exposes APIs to see the full transcript with timestamps and grab screenshots of what was on your screen at any of those timestamps.
Once it's created the routine, you can go to the Routines tab in Claude Code desktop app.
This also works with the CLI, Claude Cowork, Codex etc.
Youโll see your new routine there, and you can just hit Run Now to have it start running it.
Video attached shows this step by step.
The Clips apps is free and open source - Iโll link to it and the source code in the thread.
A 15-year-old dream has come true today. I started a PhD with the dream of creating a system that chants any Sanskrit shloka perfectly.
And here I am opening sourcing ๐๐๐ ๐๐ก๐๐ง๐ฎ - ๐ ๐ฏแน๐ญ๐ญ๐ (๐ฆ๐๐ญ๐๐ซ) ๐๐ฐ๐๐ซ๐ ล๐ฅ๐จ๐ค๐-๐ญ๐จ-๐๐ก๐๐ง๐ญ ๐ญ๐๐ฑ๐ญ-๐ญ๐จ-๐ฌ๐ฉ๐๐๐๐ก (TTS) ๐ฌ๐ฒ๐ฌ๐ญ๐๐ฆ ๐๐จ๐ซ ๐๐๐ง๐ฌ๐ค๐ซ๐ข๐ญ. This is the world's first vrutta-aware, open-source TTS for Sanskrit Chanting.
Google Flights shows you what airlines want you to see.
Claude shows you what they are hiding.
$890 flight.
Paid $97.
$793 saved in 4 minutes flat.
Here are the 7 prompts that made it happen:๏ฟผ๐
Grok 4.5, based on our 1.5T V9 foundation model, with Cursor data added in supplemental training, is now in private beta at SpaceX & Tesla. Early evals show performance close to, perhaps exceeding Opus.
RL is continuing to significantly improve the model, and the Grok Build harness gets better every day.
Nice work by all those involved!
Completely trained from scratch new models will be released by @SpaceX every month this year.
๐จ NOW YOU RUN A COMPANY WITH ZERO EMPLOYEES
Paperclip is a 100% open-source framework (70k+ stars) that makes this possible.
Rather than just prompting a model, you hire a CEO, engineers, and a QA reviewer.
Every worker is an AI agent, and Paperclip is the Node.js and React control plane that keeps them aligned.
Stop chaining messy scripts together and build a living organization:
โ Stand up a CEO agent to set strategy
โ Hire engineers and designers via Claude or Codex
โ Build in an automated QA loop before any ticket closes
โ Manage the entire portfolio from your phone
When an agent slips, you do not rewrite your whole pipeline: you just correct its persona prompt, exactly like coaching a junior hire.
It is exactly the kind of tooling the space needs right now.
Free, open-source, and self-hosted.
Repo link in ๐งตโ
This is whatโs causing Anthropic to aggressively beg for govt protection (see below). Customers are finding cheaper alternatives. Keeping employees requires continuing ultra-rich secondaries ($$$) that are dependent on revenue growth. When you canโt win on the field go to DC.
Kian Katanforoosh, Stanford CS lecturer (Forbes 30 Under 30):
"Two Sigma pays $650K a year to researchers who can train the neural networks in this lecture. I teach them at Stanford for the price of clicking play."
this free stanford lecture holds the entire "neural networks to win every trade" framework the 2026 quant threads sell you. katanforoosh stands at the board and builds a network from scratch, then lands on the one honest idea in the whole post: a neural network doesn't predict the future, it learns the expected value of an outcome given your inputs. that's it. the rest is engineering.
everything the thread codes up, the LSTM, gradient descent, the universal approximation theorem, is lecture ten of a public stanford course. the math was never a $55k secret. it's standard material, taught openly by the people who built the field.
so the framework was never the moat. a 1.3-million-view thread is reselling a lecture you can watch tonight.
and here's the honest catch the thread half-admits and then buries under "win every trade." a network only learns the right expectation if the data distribution holds still, and markets never do. the model assumes a stable world; a market shifts under it the moment you deploy. "win every trade" is the exact thing this math cannot promise. the lecture is free. knowing when your model has quietly started learning the wrong distribution is the part that actually pays $650k.
This "Taste" Skill is cracked.
I can't believe I didn't discover it sooner - everyone should install this.
It works directly inside Claude Code, Codex, Hermes & more to completely kill AI-generated slop.
If you send this prompt to your agent, it will automatically install it:
Raphael Townshend, Stanford AI PhD and founder of Atomic AI (Forbes 30 Under 30):
""Wall Street will pay you $500K a year to build these models. I'd rather teach them to you for free."
this free stanford lecture holds the entire "77% win rate, pure math" random forest the 2026 quant threads sell you. and the guy teaching it didn't take the wall street money either, townshend went on to found an ai drug-discovery company and land forbes 30 under 30.
at the board he builds it from scratch: one decision tree overfits, so you grow hundreds on random subsets of the data and features and average them. the errors cancel, the signal survives. that's the whole "100 ai agents auditing the market" idea, minus the marketing.
the โN feature rule, the out-of-bag error, the probability output, all of it is standard ensemble learning, taught free by stanford for years. random forests came out of leo breiman's public paper in 2001. the thread didn't discover it. it renamed it.
and here's the honest part the win rate hides. a model that scored 77% on past data is describing the past, not promising the future. ensembles cut variance, they don't turn a weak edge into a real one, and markets shift under the model in ways the training set never warned about. the lecture is free. knowing whether your 77% survives out of sample and on live capital is exactly the part the post skips."
This is Polymarket influencer content dressed as technical analysis. Let me be direct.
What It Actually Is
- Random Forest with 100 trees, rebranded as "100+ AI agents auditing the market"
- Sigmoid activation on RF output โ bog-standard probability calibration. 1/(1+e^-x) is not phase 2 of a novel system, it's sklearn boilerplate
- Feature matrix: price, liquidity, volume_24h, momentum_7d, days_to_expiry โ 5 features. Input to RF. โ5 โ 2 features per tree
- Entry rule: buy when market_price โค model_prob ร 0.5 โ the "double discount"
- Exit rule: sell at model_prob ร 0.9 or 7 days before expiry
- Evaluation: Sharpe ratio on log returns
What's Missing
77%/80%+ win rate
โข Claim: 77%/80%+ win rate
โข Evidence: Zero backtest data, zero onchain PnL
136 signals
โข Claim: 136 signals
โข Evidence: Nowhere enumerated. The code shows 5 features
"100+ AI agents"
โข Claim: "100+ AI agents"
โข Evidence: One Random Forest model. Basic vocab inflation
Working system
โข Claim: Working system
โข Evidence: No repo, no codebase, no deployment
Indian students are DIYing a semiconductor fab at IIT Bombay.
In just 10 months they've built:
1. A DLP-based lithography machine.
2. A tube furnace to oxidise silicon.
3. A DC plasma sputter.
Total cost: โน30 lakh.
Here's a rare behind-the-scenes look at HackerFab IITB.
Google released Study Notebooks in the Gemini app, a free interactive learning space that creates personalized lessons around each studentโs goals and uploaded materials.
Students can upload notes, syllabuses, readings, or other learning resources. Gemini then creates a diagnostic quiz to identify what they already understand and where they need more practice.
Study Notebooks can:
๐นGenerate short lessons around specific knowledge gaps.
๐นCreate practice quizzes grounded in uploaded materials.
๐นLet students pause lessons and ask follow-up questions.
๐นBreak a learning goal into more than 100 smaller objectives.
๐นTrack progress as โStrengths,โ โFocus areas,โ or โNot started.โ
๐นRecommend which lesson to complete next.
๐นTransfer sources into NotebookLM for flashcards and Video Overviews.
SAT preparation is available now using questions from The Princeton Review.
Support for JEE, NEET, ENEM, ACT, and GRE preparation is planned for this summer.
Study Notebooks are rolling out globally for free on the web in every language supported by Gemini. Mobile and school-account access will follow.
stop telling Claude Code/Codex "read this file".
stop telling Claude Code/Codex "now read that one too".
stop telling Claude Code/Codex "grep the whole repo".
install codebase-memory. it indexes the Linux kernel, 28M lines, in ๐ฏ ๐บ๐ถ๐ป๐๐๐ฒ๐. your repo takes seconds.
index once and the whole repo becomes ๐ผ๐ป๐ฒ ๐ด๐ฟ๐ฎ๐ฝ๐ต of every function, file and dependency. one query replaces dozens of grep and read cycles.
benchmarked across 31 real repos:
โ 10x fewer tokens on structural queries
โ 83% answer quality on complex tasks
โ 2.1x fewer tool calls
two prompts. send them straight to your agent ๐
Marc Andreessen says Alex Karp almost never talks about Palantir in interviews. He calls it the single best marketing strategy he has ever seen and then revealed the number that proves it works better than anything else in the history of investor communications.
Every founder makes the same mistake. They think inside out. My company, my product, my story, out into the world. It feels natural. It is also why most founder content is indistinguishable from every other founder's content.
Karp does the opposite.
He talks about the future of the US military. He talks about superintelligence. He talks about whatever is genuinely interesting to him about the world right now. And because he is the CEO of Palantir, the company just sits there attached to all of it.
Then Marc dropped the number.
What percentage of Palantir investors have read the S1? Practically zero. What percentage have seen Karp on YouTube? Close to 100.
A Edelman B2B study found that thought leadership content drives purchasing consideration more than product marketing does โ by a factor of nearly three to one among enterprise buyers. Karp did not read that report. He just built the playbook it describes.
Palantir's lawyers spent thousands of hours on the S1. It explains everything the company does with full precision. Nobody read it.
Karp spent those hours talking about things that interested him. Everybody watched.
The most effective investor communication Palantir ever produced was never filed with the SEC.
Watch the full video on @a16z YouTube channel