Practitioner's Corner
Practitioner's Corner
The Body Problem
Before an AI agent processes a single webpage, anti-bot systems have already judged it. The evaluation is about identity: TLS fingerprints, mouse acceleration curves, behavioral cadence. When the verdict is "bot," the response is often not a block. It's false data served silently. Modified prices, ghost inventory, plausible but wrong.
The agent faithfully processes poisoned input and returns a confident, well-formatted answer. The developer reads it as a reasoning failure and reaches for the standard fix: adjust the prompt, swap the model, add retries. The actual problem has no name in any failure taxonomy. That's why it persists.
The Body Problem
Before an AI agent processes a single webpage, anti-bot systems have already judged it. The evaluation is about identity: TLS fingerprints, mouse acceleration curves, behavioral cadence. When the verdict is "bot," the response is often not a block. It's false data served silently. Modified prices, ghost inventory, plausible but wrong.
The agent faithfully processes poisoned input and returns a confident, well-formatted answer. The developer reads it as a reasoning failure and reaches for the standard fix: adjust the prompt, swap the model, add retries. The actual problem has no name in any failure taxonomy. That's why it persists.
What Suchintan Singh Found When He Left ML Platforms for the Open Web
Suchintan Singh built ML infrastructure at Faire and Gopuff. Feature stores, ranking systems, search engines. Controlled environments where the data was yours and the system behaved as designed. Then he co-founded Skyvern, a browser automation company, and discovered where the engineering hours actually go. Roughly 60% of the work goes to getting past the door: anti-bot systems, CAPTCHA, 2FA, residential proxies. The agent reasoning was the part his career had already prepared him for.
What Suchintan Singh Found When He Left ML Platforms for the Open Web
Suchintan Singh built ML infrastructure at Faire and Gopuff. Feature stores, ranking systems, search engines. Controlled environments where the data was yours and the system behaved as designed. Then he co-founded Skyvern, a browser automation company, and discovered where the engineering hours actually go. Roughly 60% of the work goes to getting past the door: anti-bot systems, CAPTCHA, 2FA, residential proxies. The agent reasoning was the part his career had already prepared him for.
Talking to the Woman Who Keeps Two Hundred Robots From Getting Caught
CONTINUE READINGWhat Agents Face
Think of it as a full-body scanner. Amazon's 2026 anti-bot system evaluates six layers at once, continuously, from the first TCP SYN packet through every scroll and click: IP reputation and ASN analysis, TLS fingerprinting (JA3/JA4), browser environment detection, behavioral biometrics, CAPTCHA challenges, and ML-driven anomaly detection.
What all six layers share is an obsession with consistency. A User-Agent claiming Chrome on Windows while the TCP window size reads 29,200 and the TTL arrives at 64, both Linux kernel defaults, tells the system the session is fabricated before a single page element loads. The mismatch is the signal.
Further Reading
