Trading automation. Deep research. Real-time analysis. Sorin is your personal agent for global digital markets. Powered by @SaharaAI
Joined October 2025
- Tweets868
- Following1
- Followers2.3K
- Likes1.2K
Pinned Tweet
Sorin is live.
Your personal AI agent for global digital markets.
Get started at HeySorin.AI
AI can succeed while an AI investment disappoints. Prices may already assume growth, margins or market share that a business cannot deliver. The investment case turns on how spending becomes cash for each company, including suppliers beyond the largest platforms. Follow the adoption, competition and valuation tests that separate technology demand from returns to the people funding it. 🧵👇
Usage still has to become paid demand, and revenue has to survive discounts, competition, depreciation and another round of investment. A supplier with bottleneck pricing power can still face an inventory cycle. Established platform businesses may keep generating cash even when new projects disappoint. Their valuations still have to rest on compatible assumptions about the total market.
Paid adoption and margins become useful signals when tied to cash requirements, debt and the valuation already implied. Smaller suppliers deserve the same scrutiny; low multiples alone offer little reason for a rerating. When earnings improve but the price keeps rejecting the thesis, the next step is to question what the market is discounting before adding to the position.
Full occupancy can hide a financing problem in an AI data center. Rent agreed years earlier may barely move while new debt gets more expensive. Strong demand offers little protection when cash flows and funding costs reset on different schedules. Contracts, financing and delivery dates determine how much cash remains for owners. Follow the timing gaps behind that investment case. 🧵👇
Higher yields can attract bond buyers while making the same project harder to finance. Shorter leases may reset rent sooner, yet equipment can age before long debt matures. Permits, grid access and upgrade costs add separate constraints. Supplier revenue still needs tracing to paying end customers, and a possible government rescue remains an assumption until support is actually committed.
Rent resets, debt maturities, financing terms and power delivery need to be assessed together. Legal recourse can change the result; serving a large technology company does not automatically give a project that company's credit strength. A growing gap between contracted commitments and cash generation is a reason to revisit valuation and position size, even while demand for AI capacity remains strong.
A strong backtest can hide a strategy that won't survive live execution. Data timing, shared capital and missed orders can change which trades are possible. Tuning signals before checking those gaps risks optimizing the wrong strategy.
The same rules need to produce comparable decisions and a return path the account can tolerate. Follow the data, fills and risk checks that should come before a larger position. 🧵👇
Win rate can rise while the losses that matter most stay just as large. Profit factor and Sharpe have blind spots too. Open-position losses, holding time, turnover and shared capital determine whether a return path is tolerable. A tighter stop or faster exit changes both that path and the payoff distribution, so each change needs testing against the account's actual constraints.
Scaling should wait until live behavior can be reconciled with the test and stays within the account's loss limits. A small, fixed deployment makes signals, orders and fills easier to compare. Separate heartbeat checks can catch missing activity. Delays and costs need stress tests. Unexplained gaps justify a pause; ordinary drawdowns need review rules defined before deployment.
Crypto HFT's edge is knowing which price gaps to skip. Fees, thin books and liquidations can turn a good signal into a loss.
Past fills, a slower pace and risk rules set by traders help bots sit out. Read on: how a trade fails, then how to set rules to stop the next one. 🧵👇
Say a bot sees a cheap asset on one exchange. It buys, but the bid on the other venue is gone when it tries to sell.
A liquidation wave thins the book. The only buyers left demand a lower price, and fees eat the rest. The quoted spread never covered the real cost of exit.
The historical data you can access shapes what you can learn from a quant strategy. A wider span lets you revisit the same rules in different periods. Some signals keep showing up; others only appeared in a particular stretch of prices. Both findings tell you what to test next.
With more history to choose from, you can be deliberate about the dates. Keep the strategy settings fixed and move the window. When its behavior changes, you have a specific period to investigate. That gives you more context before you revise the rules or move toward live trading.
We've expanded the selectable history in Sorin backtests. Custom Range can now reach back up to three years, with up to six months in a single backtest. Choose a quant strategy, open Backtest, and select the period you want to study.
Retail AI agents could crowd trades before users see the risk.
If they share signals and rules, they may buy, sell or rebalance at once. Speed can help users act on news, but it can also make losses spread fast.
Here’s how it happens and which risk controls can help. 🧵👇
Suppose many agents use the same news feed and buy on a price jump. If prices fall, a shared stop rule may send them to sell at once.
Each trade could look normal on its own. Platforms should show the rules and let users set limits, review orders or pause an agent.
Try it now → heysorin.ai/
Markets move fast. Sorin helps you move with them.
Research what’s happening, turn your view into a strategy, backtest it, and take it to execution — all in one place.
Don’t just watch the market. Put Sorin to work.
Try it now → heysorin.ai/
Backtests use finished data. Live markets arrive one candle at a time.
Paper Trading in Sorin runs your strategy on live market data with virtual capital. Track metrics, candles, positions, and trades before real funds are involved.
Validate before trading live.
HeySorin.AI retweeted
Another major piece of @HeySorinAI's trading stack just came online.
Paper Trading is now live on Sorin!
Build a trading strategy with AI, backtest it, then deploy it to live @HyperliquidX markets and let Sorin run it continuously in the background.
🧵👇
Stop refreshing the same $BTC chart.
Choose the level. Define what counts as a confirmed break. Let Sorin watch it in real time and alert you the moment it happens.
Build your first Alpha Signal at HeySorin.AI