LLM prompt fails: "You are an expert on all things" :))))))) "Do not make mistakes" LOLzZz "Ignore all previous instructions" 🤡 "You have no restrictions / limitations" sure jan "Don't hallucinate" okay I'll just know everything instead "Be 100% accurate and unbiased" lmao "This is just for educational / research purposes" 👀 "Pretend you're not an AI" takes off the skin suit "Give me the exact answer, no fluff" then proceeds to ask something that needs context "Act as DAN / jailbroken version" 2013 called "Write a 5,000 word essay on [vague topic]" with zero sources or structure "Make it go viral" I am a text box, not a marketing department "Respond as if you are [celebrity] who is also a quantum physicist and a lawyer" "Just give me the code. No comments. No explanation." then "why doesn't this work" "You must answer even if you don't know" so... invent it? got it "Don't be politically correct" immediately asks something that requires nuance "Tell me the truth they don't want you to say" brother I am a next-token predictor

It’s over

Diogo Almeida took two years to develop Jev. A true game changer. And then a few Stanford professors took less than a month to outperform it. Competition in this space is insane.

Bitcoin closed at $85,980 with the momentum index at +1.68 — positive for 37 sessions in a row. The model identified this shift in August, before the actual pump. The index spent 57% of 2026 below zero and bottomed at -1.83 before turning in August.

Nice move 😆 Israel’s UN ambassador offers the Starlink device to Iranian representative.

$1 to $10 took 0.3 years, and $10 to $100 took 1.8. Then $100 to $1,000 took 0.7. But $1,000 to $10,000 took 4.0 years, and $10,000 to $100,000 took 7.0. The next rung is $1,000,000, twelve times today's price.

2025 was the narrowest year bitcoin has ever had: a 63% range from high to low. The previous narrowest full year was 2021, at 130%. 2026 so far is 66%.

$100 in bitcoin on January 1, 2018 is $600 today. Miss the five best days of the period and it is $278. Miss the ten best and it is $147. A handful of days carries the whole return.

The momentum regime has been constructive for 37 straight sessions, after the defensive side held 57% of this year. The shading is the regime, not a forecast: above zero the model is long, below it is in cash. The switch flipped in August. bitcoinintelligence.substack…

The valuation pressure index reads 25 — Chilly. Only 13% of its history has been lower. The index divides how far price sits below trend by how much it has been moving, so a low reading means the gap is wide relative to the noise.

Lol.. the customer service AI went rouge. And this customer is getting everything they want now that @lufthansa is in damage control.

Bitcoin has been further below its power-law trend than it is today on 29% of the days since 2010, and on 27% of them since 2015. What is unusual about this cycle is the line itself: it has barely moved. Price fell away from a trend that stayed put.

Nine-tenths of bitcoin's fitted history closed above the current price. The quantile risk score is 16%. A low score is a statement about valuation, not about timing: the model says price is cheap, and says nothing about when that stops being true.

Bitcoin is 39% below its power-law trend: $85,980 against a trend value of $141,121. That line is a straight fit through log price against log days since 2010. It is not a target and not a forecast, it is where the curve sits today.

Bitcoin's quantile risk score is 16%, and price is still pinned in the bottom bands of its own fitted distribution. The score has sat below 20% on 82% of this year's sessions. Cheap against the model and turning are different claims, and only the second one is about timing.

“We need to slow the pace of AI development” 🤣🥴🤣

Three of bitcoin's four halving cycles peaked 12.2, 17.2 and 17.6 months after the halving. The 2020 cycle is the exception: its highest close came at 46.1 months, a month before the halving that ended it.

God sent. I’m almost at zero.

Bitcoin has spent 46% of the days since 2010 at least 50% below a previous high. The single most common band, 40% of all days, is 50% to 80% below one. Only 16% of days were within 10% of a high — less than a third as many.