The Morning Bell No. 1: Bitcoin Broke $80K While I Was Asleep

Branden Bell
3 days ago
4 min read

Bitcoin crossed $80,000 this morning for the first time in 113 days. I was asleep when it happened. The AI trading desk I built was not, and by 8:45 it had the whole story graded, sourced, and waiting for me.

Welcome to The Morning Bell, a daily premarket newsletter I publish with my agentic AI research system every weekday. This first edition is going out after lunch because I spent the morning wiring the pipeline. Starting tomorrow it publishes before the opening bell.

I'm not a financial advisor and nothing in The Morning Bell is financial advice. This is entertainment and education based on research my AI system produces and things I may or may not be doing personally. The account featured here is a small agentic test account, fully separate from my personal trading. Always do your own research.

What you're reading

I run a research system built on Claude that connects to my Robinhood market data, my Gmail newsletters, and a forecast ledger in Google Drive. Every morning around 8 it runs my saved stock scans, reads the overnight newsletters, digests a list of YouTube analysts, checks the economic calendar, grades yesterday's ideas against real prices, and writes me a brief.

The Morning Bell is the public version of that brief.

The test account it researches for holds $117.24. That's not a typo. I want these ideas judged on their track record instead of anybody's stack size, and a tiny account keeps every claim honest. The system grades its own picks daily and I'll publish the grades here, wins and losses both.

The tape before the bell

As of 8:30 this morning, premarket:

S&P 500 (SPY): +0.5%. Bouncing after Monday's slide.
Nasdaq 100 (QQQ): +0.9%. Reclaiming the level it broke Monday.
Small caps and bonds: quiet. IWM +0.7%, TLT slightly bid, volatility falling.
Bitcoin: $80,580 overnight. First $80K print in 113 days, with $338 million of ETF inflows in a single day per Mando Minutes.

My system's regime call is amber, leaning constructive. Monday's close tripped one of its defined caution triggers on the Nasdaq, but volatility kept falling and Tuesday's premarket reclaimed the level broadly. It's not green either. Prediction markets on Kalshi priced 60% odds of a Fed rate hike this year, and Nvidia walks into Wednesday's earnings on its worst losing streak since 2022.

What the system flagged overnight

Dick's Sporting Goods missed big. Earnings came in at $3.53 against a $3.88 estimate and the stock dropped 17% premarket. My scanner caught it as the heaviest options flow of the morning.
Tesla put the Cybercab on the calendar. September 3 in Austin, per Electrek. That's a dated catalyst nine days out.
InvestAnswers is deploying, not just talking. The analyst my system weights highest bought Tesla LEAPS and Bitcoin ETF LEAPS this weekend and told viewers to stop selling covered calls on their highest-upside names. When someone's actions match their words, my ledger scores it differently.
The contradiction of the week. Kalshi says 60% odds of a hike this year while a Bank of America survey says 69% of fund managers expect a neutral new Fed Chair on Friday. Both can't be right, and Friday at 10 AM settles it.

The setups on my screen

Research notes, not recommendations, graded here tomorrow:

Momentum long, AMD. Yesterday's pullback pick from my scanner, up 3.3% premarket. Event risk is Nvidia's Wednesday report, which moves the whole semiconductor complex.
Research short, Strategy (MSTR). It raised $2 billion last week and bought zero bitcoin for the third straight week. Working so far, but bitcoin at $80K argues against pressing it.
Options study, Nvidia. An August 28 defined-risk call spread priced at about $2.26 off real chain data, risking only the debit into the earnings print, with the chain's own math giving it under a 26% chance of max profit.
Pullback watch, Astera Labs. Cooled off while holding its trend, and the analyst ledger has an accumulate level far below spot. Patience trade.

Yesterday's report card: three of four ideas moved the right way. The account traded none of them. More on that below.

Today's calendar, all times ET

Today 10:00. New Home Sales, Consumer Confidence, Richmond Fed. Intuit and Zoom report after the close.
Wednesday 8:30. GDP, Core PCE, and Durable Goods land at the same moment. Nvidia reports after the close. Biggest day of the week.
Thursday. Jackson Hole opens, and this year's theme is payments and financial innovation.
Friday 9:45 and 10:00. Chicago PMI, then the new Fed Chair's debut keynote. If you hold anything rate-sensitive, Friday morning is the moment that matters.

The lesson my own system taught me

My grader stung me this week. Last week's calls scored two right, one partial, one wrong. Respectable. The account executed exactly zero of them and bought a random biotech that appeared in no morning plan, then closed it flat.

The forecasting was fine. The execution was the hole. I built a system that catches my own drift within 24 hours, and it turns out the scariest report it produces is about me, not the market.

So the fix I'm testing is this newsletter. The record is public now, graded daily, and I can't quietly forget the ideas I didn't act on.

How this thing actually runs

The whole desk is a Claude scheduled task. It wakes up on its own, runs skills I wrote for scanning, scoring analysts, and building the brief, then delivers everything before I've had coffee. I wrote about how scheduled tasks work while your computer is asleep if you want the mechanics.

Building agentic systems like this is what I do at Bell Consulting Solutions. Mostly for client delivery operations rather than trading, but the architecture is the same: connected data, skills that encode judgment, a schedule, and a paper trail. If you're wondering what working with someone on this looks like, start with how to hire a Claude consultant, or just reach out directly.

What would you want an AI desk to check for you every morning at 8? Tell me. The best answer might become a section here.

One more time for the people in the back: not financial advice, entertainment and education only, based on what I may or may not be doing personally. See you before the bell tomorrow.

Sources: