Framer AI AgentsDesign and publish professional sites with AI
Promoted
Hesper Atlas
A map for the AI age. Buy/sell/hold signals.
16 followers
A map for the AI age. Buy/sell/hold signals.
16 followers
One map of the AI trade: 155 names across the whole supply chain, from chips to power, each with a daily rules-based buy/hold/sell signal, exact trade-plan prices, and a public track record that includes every losing signal. Educational tool.
Hey PH, solo founder here.
The premise: the AI buildout is where returns concentrate for the next several years,
so I hand-picked 155 names that trace the whole supply chain, substrates, chips, networking, servers, power,
cloud, and gave each one a systematic daily buy/hold/sell signal. Selection answers 'what to own'; the engine
answers the harder question, 'when do I get out'.
Receipts from the public ledger (closed trades, split-adjusted): it bought Super Micro at $4.25 in July 2022,
before ChatGPT existed, sold at $71 two years later (+1,574%), and was out before the -85% collapse. Vertiv,
the company cooling the datacenters, from $13.59 to $102.60 (+656%). Credo, AI networking, bought in
January 2023 a year before the story broke (+318%). Micron at $60 in March 2023 (+93%). BTC on held-out
data: +34.5%/yr vs +7.6% holding, with a third of the worst drop.
The twist that makes this different: every signal the engine produces lands on a public ledger, losers included,
regenerated nightly. The losing trades sit on the same page as the winners (SMCI's recent -9.3% is right there),
so no cherry-picking is possible.
And the claim I defend hardest is the boring one: across all 152 evaluated
names, half the worst drawdown of buy-and-hold at comparable returns.
Free tier shows three full stock pages a
day, no card.
Would love your skepticism more than your upvotes.
@hesper_atlas Upvoted. I like that you've made the signals fully transparent instead of only highlighting the winners. That makes it much easier to evaluate the system over time. Congratulations on the launch. How are you handling portfolio rebalancing and risk management when multiple signals change at the same time?
@spotch_founder Thanks, that means a lot. The full ledger is the whole point: judge it on everything, not a highlight reel.
Good question. Simultaneous changes are actually the normal case here, because everything is computed in one daily pass on closing data. Nothing reacts intraday. The model portfolio is rebuilt from that pass each day: rank the universe, take the top names, weight them by risk parity. So if three names flip on the same day, the weights re-solve as one problem instead of three trades fighting each other.
On risk: no position over 25%, correlated names get sized down together so a cluster flipping at once can't pile into one exposure, and a cash buffer grows when the basket gets choppy.
Honest caveat: we tested fancier rebalancing (regime switching, no-trade bands, factor weights) and net of costs none of it beat simple risk parity, so we shipped the simple version.
You caught a gap, by the way. This is only explained inside the app right now, so I'm adding it to the public methodology page. Thanks for the nudge.
would love to see a watchlist feature so i can flag a handful of tickers and get a single daily summary instead of scanning 155 names every morning. could be a simple toggle in settings, with the signal, plan prices, and any change from prior day rolled into one email or dashboard widget. would make it way easier to actually act on the signals without missing moves across the chain.
@hafizezdztabm Yes great idea, it's already possible to favorite tickers, but will extend the feature to have a clean watchlist
love that you publish the losing signals alongside the winners. basically most trading products hide the bad calls but having that full public track record is the kind of transparency that makes this feel actually trustworthy.
@hira51904067906 That's the goal. Transparency and killer engines. So everyone benefits, nothing hidden.