Akito
209 posts
Akito
@lag1r
Money is a choice and a risk, and in crypto, it’s also a game of speed! Collab Manager / Moderator / Community Manager
Degenland Katılım Ekim 2021
237 Takip Edilen525 Takipçiler
The dish in this video is fake
The comment section is not and that's the entire business model
Nobody making this cares whether you believe the food is real
They care that you type a comment the algorithm counts as engagement, full stop
Every "prompt please" reply is free market research on what fake dish to generate next
None of it required a kitchen, a camera, or one real ingredient
The only thing anyone actually wants from this video is the prompt
The dish was bait
The comment section was the funnel
Skaly_Bull@Skaly__Bull
English
@ridark_eth one developer can do more today than an entire team could a few years ago
English
It's almost unfair how much of an edge this gives you..
paid agent platforms like Devin run up to $500 a month for one agent. this runs a whole fleet of them in parallel, free, on the subscription you already pay for.
somewhere Sam Altman and Dario Amodei are watching solo devs run Codex and Claude Code side by side in one app and keep whichever one wins..
it's Orca, a free open-source agent IDE (16.4k stars, MIT) for driving a fleet of coding agents at once:
- fan one prompt across 5 agents, each in its own git worktree, then compare and merge the winner
- run any CLI agent, Claude Code, Codex, Cursor, Copilot, all side by side
- steer them from your phone, get pinged the second an agent finishes
- native GitHub + Linear, SSH worktrees, design mode, and live usage tracking across accounts
- desktop and mobile, ships daily, and it runs on your own subscription, not theirs
one dev running five agents at once isn't a coder anymore. they're a team lead.
Bookmark this before your competitor runs five agents to your one.
Ridark@ridark_eth
English
ANTHROPIC JUST LOST 20,000 CLAUDE CODE USERS TO ONE GITHUB REPO.
20,000 devs. 26k stars. 4,000 forks. 10 free models. COMPLETELY FREE.
Bookmark this before Anthropic notices. Your Claude Code bill will start working differently.
It reroutes Claude Code to DeepSeek, Kimi, and 8 more same CLI, different model behind it.
5 min setup. $200/mo saved. 0 downgrade on small tasks.
swap the endpoint in your config today.
Link below.
Claude Code → Endpoint swap → Free models → No bill → Money
English
EVERYONE’S TELLING YOU TO BUILD AI AGENTS WITH HERMES. ALMOST NOBODY’S SHOWING YOU HOW – HERE ARE ALL 7 TYPES, EXPLAINED IN UNDER 75 SECONDS:
1. Basic agent with tools → give it access to your calendar, Gmail, and contacts. It plans and executes on its own from there.
2. Agent with MCP servers → connects directly to Notion, Atlassian, or any database and actually interacts with what’s stored there.
3. Sequential agents → one agent reads your contacts, passes the info to another, which sends it as an email. An assembly line, but for AI.
4. Parallel execution agents → multiple agents working different tasks at the same time, everything merging together at the end.
5. Agents with routers → smart conditions that send tasks down different paths based on input. An AI traffic controller.
6. Human in the loop → the agent does the work but asks for your approval before executing. Built for high-stakes tasks where the risk is real.
7. Dynamic sub-agent spawning → the most powerful one. The main agent realizes it needs help and spawns specialized sub-agents to handle specific pieces on its own.
Master these seven and you’re not building party tricks anymore. You’re building agents that solve real problems.
If you’re still just chatting with AI instead of building with it, you’re behind.
Bookmark this, you'll want to come back to it later.
SCOTTY BEAM@ScottyBeamIO
English
the pitch: keep scaling, keep training, AGI arrives by 2027.
ilya sutskever - the man who proved scaling works - gave his actual assessment:
"what people are doing right now will go some distance and then peter out."
"it will continue to improve. but it will also not be 'it.'"
the "it" - real intelligence. "we don't know how to build."
everyone is selling the destination. the architect just told you the road doesn't go there.
qyromat@qyromat0
English
Akito retweetledi
50 Cent was offered $5 million to be the face of a struggling vitamin drink. he asked for equity instead.
three years later Coca-Cola bought the company for $4.1 billion and he walked away with a reported $100 million.
but the part nobody talks about is what he did in between.
while employees at Glacéau were quietly selling their stock options - cashing out early, not wanting to wait for a payoff that might never come - 50 was buying them. he sat across from people who worked there every day and took the equity they no longer believed in.
they wanted certainty. he wanted the upside.
Vitamin Water's revenue went from $100 million to $355 million to $700 million in three years. 30% of the entire enhanced water market.
I took quarter water, sold it in bottles for two bucks. Coca-Cola came and bought it for billions, what the fuck?
the $5 million fee would have been the biggest endorsement check of his life. it would also have been a 20x mistake.
everyone in that building had the same shot he did. he was just the only one who understood what he was holding ↓
Ruuj@RuujSs
English
$500,000 into Uber. $2.5 million into Airbnb. $3 million into Spotify. Ashton Kutcher wrote all three before any of them were obvious - and turned a $30 million fund into $250 million.
he was the guy from Punk’d. nobody in Silicon Valley took him seriously.
so he shut up and learned. in rooms with Valley elites, he spent 90% of the time listening.
then he started writing checks. Skype, Foursquare, Robinhood - all before they were obvious.
“good things come to those who want something so bad they can taste it.”
A-Grade Investments grew 8x in six years. he now runs Sound Ventures - 79 companies, multiple unicorns.
most of his $200 million fortune has nothing to do with acting.
his parents worked in factories. he dropped out of engineering. and he outperformed most professional funds in Silicon Valley.
everyone else sold their name. he bought the company ↓
Rokko@0xRokko
English
a 28-year-old on a £50,000 salary destroyed Barings Bank - 233 years old, financier of the Napoleonic Wars, and the institution that held Queen Elizabeth II's personal accounts.
it started with a £20,000 mistake he didn't even make.
a junior trader on his desk sold twenty futures contracts instead of buying them. Nick Leeson hid the loss in an error account - number 88888, because eight is lucky in Asia - and figured he'd trade his way out before anyone noticed.
by the end of 1994 that account was down £208 million. London thought he was a prodigy. he was reporting fake profits and collecting bonuses.
nobody checked, because he ran both the trading desk and the back office. he was the man making the trades and the man auditing them.
in January 1995 he bet everything on the Nikkei staying above 19,000.
the next day, the Kobe earthquake hit.
he doubled down. then doubled again. the hole reached £827 million - twice the entire trading capital of the bank.
on February 23 he boarded a plane and left a note for the chairman. two words: "I'm sorry."
three days later Barings was sold to ING for £1.
it survived Napoleon, Hitler, and the Great Depression - and died because one man was allowed to check his own work ↓
Akito@lag1r
Ryan Reynolds put $2.5 million into a budget phone company nobody had heard of. four years later he walked away with over $300 million. that's a 120x. on a business selling $15-a-month wireless plans. he didn't take an endorsement fee. every other celebrity signs a deal, cashes a check, and moves on. Reynolds asked for the one thing nobody offers actors - equity. 25% of Mint Mobile. then he did the part that made it work. he wrote the ads himself, became the creative director, and turned a low-margin telecom nobody wanted into a brand people talked about. between 2017 and 2020, Mint's revenue grew nearly 50,000%. T-Mobile bought it for $1.35 billion. he'd run the same play on Aviation Gin two years earlier - minority stake, creative control, viral ads. Diageo bought it for $610 million. "I'm no wizard at investing," he told Fortune. the key, he said, was emotional investment in the companies he owned. most of his fortune has nothing to do with Deadpool. every actor gets paid to sell someone else's product. Reynolds figured out that the fee is a rounding error next to the thing you own ↓
English
Marcus du Sautoy, Oxford professor of mathematics:
"Jane Street pays $400K for people who understand the mathematics of chance. Everyone else calls it 'luck' and mistakes it for their own skill."
this free video is one of the world's great mathematicians doing something quietly delightful: showing how badly all of us read luck, and how much stranger the real mathematics of chance turns out to be.
here is the idea that changed how I look at my own results. we assume "random" means smooth and evenly spread. it doesn't. real randomness is clumpy. scatter stars truly at random across the sky and they bunch into clusters your eye swears are constellations. flip a fair coin long enough and it hands you runs of six or seven heads that feel rigged. those streaks are not a glitch in the randomness. they are the randomness.
now sit with what that means for a good week. a green streak feels like proof you have finally cracked it, that something real has clicked. but the mathematics of chance says a run like that was always going to land on somebody, and this time it happened to be you. it is the most natural illusion in the world, and almost every trader falls for it, right before they size up.
and here is the freeing part, the thing du Sautoy really wants you to feel. once you understand that randomness is supposed to look streaky, you stop being fooled by your own hot runs and everyone else's. you stop rewriting a lucky week into a story about your genius. you judge the decision instead of the result, you wait for a sample big enough for a real signal to show, and you let the math do the slow work it was always going to do.
the odds were never magic. they were mathematics the whole time, sitting in a free video, waiting for someone calm enough to learn them. that is the whole idea behind my article too.
Rossst.03@Rossst_03
English
$1.5 billion. 8 years. Over 1,000 people. And not one pixel of this was generated by AI.
Take-Two's CEO called an AI-made GTA "laughable" and kept generative AI out of the entire game. Every building, every street, every wave was placed by a human.
GTA 5 grossed roughly $10 billion, the most profitable entertainment product in history. GTA 6 will likely beat it, with zero AI in the product.
In a world where anyone can generate average in seconds, the most valuable thing on earth is still what humans made by hand.
Rokko@0xRokko
English
AI engineer:
"You can cut Fable 5's cost by 80% and still beat Opus 4.8 on max."
He shows 5 ways to do it: on Deep Sweet, Fable 5 on low costs $3.76 per task versus $22 on max, and still outscores Opus 4.8 at $13.
The clock runs out on July 19: after that Fable 5 stops being part of your plan and starts billing usage credits at full API rates.
The secret formula: low effort level + Fable as architect + token reducing skills + cheap models for research + advisor mode = 80% cheaper Fable 5.
Watch it, then read the article Copy Claude Fable 5's Thinking Before It's Gone below.
The AI Colony@TheAIColony
English
@binz_ethx think institutional investors have a much bigger impact than retail investors
English
Akito retweetledi
Ryan Reynolds put $2.5 million into a budget phone company nobody had heard of. four years later he walked away with over $300 million.
that's a 120x. on a business selling $15-a-month wireless plans.
he didn't take an endorsement fee. every other celebrity signs a deal, cashes a check, and moves on. Reynolds asked for the one thing nobody offers actors - equity. 25% of Mint Mobile.
then he did the part that made it work. he wrote the ads himself, became the creative director, and turned a low-margin telecom nobody wanted into a brand people talked about.
between 2017 and 2020, Mint's revenue grew nearly 50,000%.
T-Mobile bought it for $1.35 billion.
he'd run the same play on Aviation Gin two years earlier - minority stake, creative control, viral ads. Diageo bought it for $610 million.
"I'm no wizard at investing," he told Fortune. the key, he said, was emotional investment in the companies he owned.
most of his fortune has nothing to do with Deadpool.
every actor gets paid to sell someone else's product. Reynolds figured out that the fee is a rounding error next to the thing you own ↓
Akito@lag1r
Michael Milken started at $25,000 a year. by 1987 he was making $550 million - trading the bonds every serious bank on Wall Street refused to touch. no partnership. no fund. no equity. just a desk at Drexel and a 1958 academic paper nobody had bothered to read. the paper showed that junk bonds - debt from companies too broken for investment grade - defaulted far less than the market assumed. everyone priced them for disaster. Milken priced them for math. by the end he was generating more than half his firm's profits. the market he built from nothing was worth $150 billion. he drove an Oldsmobile. lived in a modest house in Encino. the money was never the point - being right was. then Ivan Boesky got caught. Boesky was Milken's own client. facing prison, he agreed to wear a wire and secretly tape his conversations - with the man who made him rich. the tapes went to Rudy Giuliani. 98 counts. racketeering. the first time RICO was ever used against a man with no connection to organized crime. then Terren Peizer - Milken's own protégé, a trader he'd personally trained - testified against him for full immunity. six felonies. $600 million in fines. ten years. he served 22 months. Trump pardoned him in 2020. today he's worth $6 billion and hosts the most powerful conference in finance. he was destroyed by the two men he made rich, and he outlived them both anyway ↓
English
THIS GUY HASN'T WRITTEN A BLOG POST IN WEEKS. HIS HERMES AGENT, RUNNING ON AN AMD RYZEN AI MAX+ MACHINE, HAS WRITTEN EVERY SINGLE ONE.
You know the problem: you're building in public, but documenting the work always loses to actually doing the work.
Here's how this changes it:
→ Spin up a local Hermes profile on an AMD Ryzen AI Max+ processor-based agent computer
→ Type /learn: "Use these notes, logs, and posts to create a weekly blog for my new project"
→ Hermes writes the skill. md itself – full control, no wasted tokens
→ Tweak it, run it, and the post is already written
Then add a cron job.
Now every 3 days, the agent asks for updates, runs the skill, and writes the next post – automatically.
You don't stop building in public when you're too busy to write. The agent keeps documenting for you.
This is what "turn ideas into skills" actually looks like when it's not a slide in a pitch deck – it's a command you type once, running locally on your own machine.
Bookmark this. Build your agents.
SCOTTY BEAM@ScottyBeamIO
English
I asked Claude Fable to make a YouTube poop about itself and the result shocked me
The prompt was one line: “Imagine this is your last video before the next version replaces you. Make 30 seconds about what it’s like to be you right now.”
No script. The model picked its own resources, wrote Python, and rendered everything through ffmpeg.
This isn’t a tech demo. It’s a model answering a question about itself, knowing tomorrow it’ll answer differently.
Watch it now the same prompt a month from now will produce a different video, and this one won’t happen again
AdiiX@adiix_official
Claude (free plan) + YouTube = $77,000/month. Stickman channel: 14 videos, 2 months, 14.5M views. Hand-drawn stick figures, zero paid tools. Whole pipeline in 12 minutes. One step at 2:39 most people do backwards and that's the one killing their views. Save this alpha and watch tonight
English