For new founders this is a great reference of what a high quality launch video looks like.
Notice he doesn't say "AI" anywhere, you know watching it is AI – he doesn't need to say it.
Also notice the simplicity. It is so simple. I could give this to an 8th grader and they would understand. Lastly, tasteful and good audio/video. Impressed.
Godfather of AI: "If you sleep well tonight, you may not have understood this lecture."
This 47-minute lecture is the best thing I saw about AI in the last few months.
It will definitely help you understand how it actually works and where it's going.
Geoffrey Hinton built the neural networks behind every AI alive, then quit Google to warn the world about it.
The part nobody wanted to hear:
> AI is already developing abilities its creators didn't intend
> in most cognitive tasks it's already ahead of us
> the question is no longer if it surpasses us but when
> the only decision left is which side of that line you're on
Right now the average person opens Claude, types something, gets an answer, closes the tab.
They think they're using AI. they're using maybe 10% of it.
I went through his entire lecture, built a practical system from what he was describing.
18 steps to actually use Claude the right way, with copy-paste prompts that work today.
Full guide in the post below.
I just uploaded a video on implementing Qwen 3.5 from scratch.
> Implement RoPE
> Implement Group-Query Attention
> Implement Recurrent Linear Attention
> Implement KV Cache Management
> Implement Decoder
checkout here:
Coding Qwen 3.5 LLM from scratch!
https://t.co/POjZ5SQ6QN
Jane Street pays $650,000 a year for quants. MIT wrote the exact bible to get there & released it for free.
51 pages. Zero to quant. Probability, stats, market making, real interview questions from Jane Street, Citadel, Two Sigma & more. Bookmark, before someone takes it down.
Dexter is an autonomous financial researcher that lives in your terminal thinks, plans, analyzes real-time market data, and self-improves as it works.
17K GitHub stars. Compatible with Claude, Codex, and OpenClaw.
AN AI HEDGE FUND ANALYST THAT GETS BETTER EVERY TIME IT RUNS.
GitHub: https://t.co/c5GeZ4KpEV
If you're a student wondering what you should study in the world of AI, it's still the same: math, physics, chemistry, biology, computer science, engineering.
STEM teaches reasoning, objectivity and how to think. It makes you better at learning everything else. In STEM, ideas win, not institutions.
There will always be career opportunities for the analytically gifted in these fields. It makes you better at everything else. It's hard, it's competitive and it will hurt your spirits to grasp things, but that's when you uplevel yourself. And if you get used to constantly upleveling yourself, you will do quite well and be quite satisfied.
Even today, the best AI models are reinforcement learned from human experts. If you build, you can keep pushing the needle.
a Bridgewater intern spilled his entire strategy on a Royal Caribbean deck and didn't realize it
day three of a mediterranean cruise. i was reading on the upper deck. he sat next to me and started a phone call.
no headphones. speaker low but not low enough.
"the model is simple. we don't predict outcomes. we filter wallets by disposition ratio and only copy their dominant category. exit on volume spike or 85% target hit."
i put my book down.
he kept going. twenty minutes. explaining to someone on the other end exactly how the fund's prediction market desk works.
"win rate means nothing. a wallet winning 85% can lose money. what matters is how much of each winner they capture vs how much of each loser they hold. we target 0.70 minimum"
"every wallet has one strong sector. politics. crypto. weather. you copy that sector only. copy everything and you import their losses"
"capital velocity is the multiplier. best wallets recycle each dollar 49 times. average does 6. same edge per trade. 8x more cycles"
he hung up. looked at the sea. never looked at me once.
i went to my cabin. hands shaking. opened Claude.
typed everything word for word.
"build a copy engine using these four filters. disposition ratio above 0.70. dominant category only. kelly sizing. exit on 3x volume spike or 85% of expected move"
Claude needed an execution layer. gave it:
https://t.co/GvslFlf9iR
1,1k+ stars. official Polymarket SDK. price feeds. order placement. WebSocket streams.
the terminal was live before we docked in athens.
by the time i got home the P&L was already at +$1,400. i hadn't opened the laptop once.
today: +$12,685. 87% win rate. 34 markets traded.
the terminal runs 6 strategies at once. base rate. conditional probability. liquidity snipe. delta hedge. calibration. theta decay.
every entry scored on disposition ratio before execution. anything below 0.70 gets killed automatically.
AMD Xilinx - entered 52c. model 59c. +7c in 2h40m.
BTC 92K Oct - entered 41c. model 42c. +1c in 1d2h.
Artemis launch - entered 63c. model 85c. +22c in 5h10m.
Derecho MW - entered 71c. model 87c. +16c in 1h50m.
spread arbi catching cross-platform gaps live:
BTC 92K vs Kalshi gap 0.3c
Fed cut vs PredictIt gap 3.7c
AAPL>180 vs Manifold gap 9.3c
Derecho vs Betfair gap 11.9c
flow pressure: 67% buy. 33% sell.
oracle delta tracking NOAA, AP, CoinGecko, Reuters, Stochastic feeds. latency under 130ms on all six nodes.
copytrade here: https://t.co/PTZuvexxtE
total cost: Claude plus a VPS. less than one drink on that ship.
i never spoke to him. never introduced myself.
he has no idea someone was listening.
and now thousands of you are reading what he said.
sorry about that.
yo viendo a usuarios de Claude automatizar negocios enteros
sacando $10k, $50k, $100k al mes, mientras yo sigo escribiendo prompts a mano
y alguien soltó una guía completa de 4 horas sobre cómo lo hacen
mi reacción:
built “Palantir for Family Trips” for our Yosemite weekend with friends
it has:
- route playback and convoy tracking
- meal logistics
- family readiness + expense tracking, and
- a full ops timeline for mission success
absolutely unnecessary, but worked surprisingly well
(repo 🔗below)
Constantly dope your mind with reading :
Read a lot of
- Human Evolution
- Physics
- Metaphysics
- Mathematics
- Psychology
- Science
- History
- Civilizations
- New Technology
- Geopolitics
- Great Biographies
- Elon Musk urgency
- Robotics & Ai
You can grow your intelligence
You can grow your intelligence
You can grow your intelligence
You can grow your intelligence
You can grow your intelligence
Tim quant dari BlackRock, menggunakan machine learning dan AI berhasil mengalahkan market dalam periode 5 tahun terakhir (about 94% of the time)
Saat ini, kelompok tersebut mencakup 125 orang dengan gelar PhD atau S2, termasuk setidaknya satu ilmuwan roket yang pernah bekerja di NASA
Kira-kira orang-orang seperti itulah yang kita hadapi saat memasuki stock market US. Gimana pede berani lawan?😁
Sumber: https://t.co/F0V3ooqTbm
Here’s what you should actually do if you suddenly run up a large amount of money:
You will most likely never see this much money again
Don’t fuck it up
The interest alone, or a few years of compounding is worth far more than you realise. If you’re lucky, you can live off interest alone. $3,000,000 at 5% a year is 150k PA. Even getting to half that amount is 75k PA for literally doing nothing
Lock the majority of the money down first (high yield savings account, no withdrawals, no impulsive spending). Put tax money aside immediately
Clear out every high interest debt and don’t take on new ones. People with 500k net worth are asking me if they should spend 100k on a car. No. Do not do this. I absolutely love supercars, but I drive a Kia Sportage
Build 6–12 months of emergency savings
Buy a home if the numbers make sense (they usually do)
Build a real budget: calculate your annual costs, your net worth, and how much interest you generate each year, are you spending more than you’re earning? Track your costs monthly so you always know where you stand
Once you're grounded put the bulk into long term investments (ETFs, super, index funds)
Keep a small liquid stack for opportunities
Don’t let lifestyle creep drain your future. Just do what you usually do. Clout is worthless
Set up proper tax and wealth structures (trust, company). See a financial advisor for this. They also know how to reduce taxes in ways you never thought possible. Protect yourself, insurance, wills, estate planning, legal structure
Build a long term financial plan and actually stick to it
I'm sure there's a lot more you could add to this list, and please let me know. Being a responsible adult isn’t that glamorous but it keeps you from going back to work
I do not want to go back to work
- Unc
2 days until APE Season 1 begins ⏰
69M APE points over 12 weeks. Multiple boosters: Mantle deposits, DEX pioneer wallets, APEX staking, team multipliers, and early adopter bonuses.
🎁 Surprise: Trade between Sept 29 and Oct 6 to unlock an additional early adopter bonus before the official launch.
Find out more: https://t.co/AKADWEsUNz