You’re locked into a week where the tape still looks rich enough to seduce people into bad decisions, and that is exactly when the costume department in AI trading does its best work. I came back to the mic with last week’s boundary question still ringing in my ears — where the regulated product ends and the pitch begins — and this time the answer was uglier: sometimes there is no product, just a name engineered to sound like one.
The Weather Report
The broad tape is still strutting even if the floorboards creak. The S&P 500 is camped near 7,690, up roughly two and a half percent for the month, while the Nasdaq has clawed back most of July’s tech unwind. Underneath that, the Russell 2000 is quietly leaking three to four percent since August. That is not a clean all-clear. That is a market being hauled by a few big engines while the smaller boats take on water.
Rates are still the real bouncer at the door. The ten-year Treasury is around 4.3%, and fed-funds futures are putting a 64% probability on a September 16 hike. So the mood of the tape is hot, but it is hot on hope, not certainty. That matters, because when investors see mega-cap resilience and rate risk in the same frame, the appetite for a machine that promises certainty goes through the roof.
And the AI angle is the accelerant. Nvidia just posted 106% year-over-year revenue growth, and the Magnificent Seven have barely flinched through the chop. The genuine AI-infrastructure trade is holding together, which is real. But that real strength is exactly what every promo artist bolts onto the deck. A strong underlying theme does not validate the salesman waving it around like a forged backstage pass.
This Week in AI Investment Technology
So let’s go downstairs and inspect the specimens, because this week the whole scam economy showed up in one lineup.
Blue Peak Brokerage Group sounds like a real regulated firm, but the registries come back empty and the surviving Blue Peak trail leads straight into warnings, scam flags, and tip-funnel behavior.
This is Prediction 1 with the serial numbers still warm: new fake-AI trading products appear faster than anyone can shut them down. This week’s reading is brutal. Run the exact name “Blue Peak Brokerage Group” through FINRA BrokerCheck, SEC IAPD, and the cited NFA search path from the brief, and you get nothing current. Zero. SEC Investor.gov says that absence itself is a red flag. The only IAPD-adjacent BluePeak entity in the packet, BluePeak Wealth Management, Inc. (CRD 211512), shows a Georgia registration that lapsed on December 31, 2016. The regulated version of the name is not alive. The costume is.
Read the red flags →📡 The Ledger ·
On the record: Prediction 1: new fake-AI trading products appear faster than anyone can shut them down.
This week’s reading: This week’s receipt is Blue Peak: no current record under the exact broker-style name in FINRA or IAPD, while related Blue Peak entities and domains are already sitting under regulator warnings and scam flags.
Tradey Algo sells autonomous sophistication with hidden pricing and stale integration language, and that alone is enough to make the pitch wobble.
Prediction 4 says truly autonomous, no-human-in-the-loop platforms are rare and expensive. This week’s reading: when somebody says institutional-grade automation and refuses to publish the price, grip the rail. Tradey Algo’s public marketing and VSL pages, as described in the brief, do not state a subscription cost, license fee, or pricing tier before email capture or demo capture. Independent review synthesis in the brief says the software requires payment to unlock but that concrete price points are not part of the initial marketing experience. Translation: the cost reveal happens after the closer has your phone number.
Watch the VSL →📡 The Ledger ·
On the record: Prediction 4: truly autonomous no-human-in-the-loop platforms are rare, and the real ones are expensive.
This week’s reading: Tradey Algo fits the costume perfectly this week: autonomy language, no public price before the sales gate, and performance claims without public methodology.
Finimize is not the same species as the ghost brands, but its winner-card marketing still outruns the plainspoken disclaimer that tells you what it really is.
Prediction 5 says most “AI trading” apps are really recommendation feeds, not algorithms. Finimize is the clean, honest contrast case this week. It publishes pricing on the actual website: $200 a year or $29 a month, with a seven-day free trial that auto-converts unless cancelled. Its terms say, in plain English, that the platform’s content is for public information, is not tailored to an individual’s financial situation, and should not be interpreted as a recommendation to buy, sell, or hold a security. The terms also say Finimize is not regulated by the UK FCA and that users rely on the content at their own risk. Beautiful. That is what disclosure looks like when adults are in the room.
Read the terms →📡 The Ledger ·
On the record: Prediction 5: most AI trading apps are stock tips or research feeds, not autonomous algorithms.
This week’s reading: Finimize is the receipt this week: a disclosed research product with posted pricing and explicit not-investment-advice terms, not an autonomous execution engine.
Kalshi’s court fight is turning into the live-fire test of whether prediction markets are treated like financial contracts or sportsbook wagers, and the federal regulator is still watching from the parking lot.
Prediction 9 says prediction markets gain real popularity but get lumped in with sportsbooks and fought as gambling. This week the reading is screaming. On August 28, the Ninth Circuit ruled against Kalshi in No. 25-7516, holding that Nevada can enforce its gambling laws against Kalshi’s sports-event contracts. Kalshi is not going straight to the Supreme Court; according to the named reporting in the brief, it is pursuing en banc rehearing in the Ninth Circuit. Robinhood filed a cert petition on September 9. New Jersey filed its own Supreme Court petition on September 2, docketed September 8, with the state attorney general explicitly framing the issue as states’ authority to regulate sports gambling on prediction markets.
Read the Ninth Circuit opinion →📡 The Ledger ·
On the record: Prediction 9: prediction markets gain popularity but get fought as gambling.
This week’s reading: This week’s receipt is direct: the Ninth Circuit sided with Nevada, Kalshi is seeking rehearing, Robinhood and New Jersey are racing to SCOTUS, and the CFTC remains publicly silent.
Clark’s Corner
Last week I asked where non-discretionary ends and delegated management begins when the interface makes the gesture feel the same. This week the question moved again, because Finimize did something a lot of these outfits never do: it gave me a denominator. It showed at least some of the losing side, not just the hero shots.
And I still do not think that gets me to the thing I actually need.
An honest hit rate, a complete archive, every winner and every loser lined up under bright lights — all of that still tells me what a system or research product was good at finding in conditions that are already gone. The Nebius market is over. Tomorrow’s market has not happened yet. So the open question tonight is tougher than last week’s and meaner too: is there any performance disclosure format in this business that gives a buyer genuine forward-looking information, rather than a prettier autopsy of the past?
I do not know. I’m not sure the answer exists. And if it does not, then the whole category — from ghost broker to honest newsletter to courtroom knife fight — is still orbiting the same wrong question, just with different tailoring. That’s where I’m leaving it. Not closed. Just sharper.