The window to find a software flaw is shrinking to hours
Quarterly security reviews no longer keep up when flaws are found almost immediately after deployment.
Lightspeed Venture Partners · The End of Human-Based Cyber Defense | Yossi Torati, A SecurityCompanies once had months or even years before attackers found the gaps around a software update. That safety window is now shrinking to hours, and sometimes minutes.
That is the shift described by Yosi Taratti, CEO and co-founder of A Security. He pointed to signs from several sources tied to the most advanced AI models, indicating that automated attackers can spot an exploitable gap almost immediately after a change goes live.
The result is that the old routine of a security review every few months, sometimes with outside experts hired to try to break in, no longer keeps pace. Companies are now under pressure to answer the same question right after they deploy something: what did this change just make possible for an attacker?
Questions people ask
Why are attackers suddenly so much faster?
AI attackers can work continuously, without sleep or breaks, and can probe software for mistakes at machine speed. Taratti says speed is the clearest difference from human attackers, and that tools built for human pace struggle to keep up.
Can't companies just install security patches faster?
Patching helps, but it is usually too slow. Researchers need days to find a flaw, vendors need time to fix it, and large companies need more time to test and install the fix. By then, automated attackers may already have slipped through the open gap.
What are companies supposed to do instead of quarterly checkups?
Security experts now recommend continuous testing that runs every time something changes. The goal is to map the full path an attacker could take to cause real harm, and then block that path with fixes or extra safeguards right away.