Solty

My friend runs a company with no employees. Just AI agents. Last month it cleared $610k, and he was at the beach for most of it. I asked him how one person runs a whole company. He sent me the talk that made it click. A free Anthropic breakdown of how AI agents actually work, from a single prompt, to workflows, to agents that decide their own next move. I watched it last night. Halfway through I realized the hard part was never the agents. Anyone can spawn one. The hard part is the walls: a gate so they can't spend your money without a yes, a log so nothing gets lost, one source of truth so they never drift. That is the whole difference between a company and a chat window. The people who actually build the walls will be running companies while everyone else is still typing prompts into a box. Bookmark this and read the article below.

My friend applied to 150 tech jobs in two years. No MIT. No Oxford. Last month Anthropic offered him $559,000. I asked him how he does it with zero background. He sent me the exact lecture that taught him how these models actually work. A free Stanford course on diffusion and large vision models, the same tech behind every AI video and image tool people are paying for. I watched it last night. Halfway through, I realized making this stuff is nowhere near as complicated as the gurus selling $500 courses want you to think. Bookmark this. The people who actually watch it will be months ahead of everyone still guessing.

@Neuron_404 the "code of civilization is words" framing is why LLMs are the real turning point

Yuval Noah Harari, historian and author of Sapiens: "AIs are hacking the code of human civilization." In his Oxford lecture he explains why: money, laws, religion, whole states are all made of one thing - words. For thousands of years that code was ours alone. "The cows could not open a bank account. The horses could not hire a lawyer." The danger was never a robot with a gun. It's a machine that can write the stories - money, laws, beliefs - better than you, and we may end up living inside fictions no human ever wrote. Bookmark - you'll want to come back to this one.

@kaworu0x true of almost every overhyped technology, quietly

@0xSolty the original use case aged better than the hype

The man who invented the quantum computer promised it would do one small thing. Everyone who got rich off his idea quietly dropped that part. In 1981 Richard Feynman said it plainly: nature is quantum, so to simulate nature you need a quantum machine. That was the whole pitch. Not curing cancer. Not breaking encryption. Not beating the markets. All of that got added later, by people selling, not building. And the one modest job he named is still the only use experts agree is real. The founder promised the least. Everyone after him shipped a costume. Full breakdown below.

@Psalteric and it's the one job a quantum computer is actually good at

@0xSolty Nature is quantum strong definition

My friend just landed a $400k quant role. No finance degree. No Wall Street. I asked him how he broke in from zero. He sent me the exact course that got him there. A free 3-hour crash course in the machine learning strategies real quant funds actually run. Lachezar Haralampiev walks you through three full quant strategies in Python, start to finish. No fluff, no paywall. I watched it last night. Halfway through, I realized building quant strategies is embarrassingly simple. Months, not years. Bookmark this or you will lose it in the feed and never find it again. Then read the article below. • 00:00 - Algorithmic Trading & Machine Learning Fundamentals • 15:25 - Building an Unsupervised Learning Trading Strategy • 2:05:08 - Building a Twitter Sentiment Investing Strategy • 2:28:08 - Building an Intraday Strategy using the GARCH Model

@Polymarket whoever's writing these announcements has fully run out of normal ideas

JUST IN: Trump reveals the White House ballroom roof will feature one of the largest military drone ports “anywhere in the world”

An 18 year old girl just exposed what might be the biggest lie in modern science. And she proved you can do the same thing for free, on a laptop anyone owns. Investors have poured over $2 billion into the quantum dream. Famous physicists sell it from the stage. Then one undergrad, handed the field's crown jewel to confirm it was real, did the opposite. She showed the machine everyone was waiting for was never needed at all. What she found broke the hype so cleanly the whole field had to check its own fine print. Her name is Ewin Tang. In 2018 her professor, Scott Aaronson, gave her quantum computing's most celebrated result: a recommendation engine, the "you might also like" problem, that supposedly only a quantum computer could run that fast. The strongest proof anyone had that the quantum future was real. She was meant to confirm it. Instead she wrote a version that runs on a normal laptop, just as fast. No qubits. No billion-dollar machine. The famous speedup was never a speedup. Nobody had bothered to check it against clever enough ordinary code. Then she did it again. And again. Knocking over a whole row of "only quantum can do this" results the same way. Here is the part that should scare everyone in that $2 billion: most of what gets sold as "quantum AI" today is not quantum at all. It is classical math wearing a costume. Once you know how to check, you see it everywhere.

@hitu_monke best part, she was told to confirm it, not break it

@0xSolty an 18 year old just doing it on a laptop is a great story

A problem simple enough for a child sat unsolved for 80 years. The bounty for cracking it: $500, from Paul Erdős himself. Write an endless list of +1s and -1s, any pattern you like. Now add them up along every evenly spaced step, every 2nd, every 3rd, every d-th. Erdős asked: can you keep all those totals from ever drifting far out of balance? He bet you couldn't. That some spacing always blows up. It sounds trivial. It broke the best minds in the world. Hundreds of mathematicians attacked it. A huge open collaboration chipped away and stalled. It took until 2015, and Terence Tao, arguably the smartest man alive, to finally close it and collect the $500. The problems that look most trivial are hiding the deepest structure. The easy-looking ones are where the real work is buried.

@Pixel_Neuron exactly, the boring-looking ones are where the gold hides

@0xSolty Terence Tao finally closed it in 2015. The easiest-looking problems often hide the deepest structure.

@unusual_whales the domestic-only focus looks safe until it suddenly doesn't

The bank has roughly $1.9 trillion in balance sheet assets across consumer banking, commercial banking, corporate and investment banking, and wealth and investment management, and is almost entirely focused on the US. Read more: unusualwhales.com/news/wells-far…

BREAKING: Wells Fargo says it is "big time bullish on the US."

@Polymarket five years is elon's favorite number, still waiting on 2017's robotaxi

BREAKING: Elon Musk predicts AI could exceed the sum of all human intelligence in about five years.

@chooserich nostalgia hits different when the place is still standing but gone

This ain’t the city no more 😂

@crimson_moti the number is the product, that's the whole game

@0xSolty Impressive benchmark, zero practical use, quantum headlines in a nutshell.

Google's quantum computer did in 5 minutes what would take a supercomputer 10 septillion years. A 1 followed by 25 zeros. It's also completely useless. The task, random circuit sampling, was built to be hard for classical computers. It solves no real problem. Nobody needed that number. It's a drag race with no road at the finish line. Real milestone for the hardware? Yes. Proof quantum now beats your GPU on anything you actually care about? Not even close. This is the pattern behind almost every jaw-dropping quantum headline. Huge number, zero use. I wrote 5 tests to catch the trick in 10 seconds. Full breakdown below.