AivexaNewsSearch
AI news for builders and product teamsChecked every hour

The ‘Vulnpocalypse’ is coming for the internet’s aging code, says Wiz’s CTO

Collected Oct 28, 2026

Two critical vulnerabilities in PHP, the open source language that underpins roughly 70% of websites, surfaced last week and sent organizations scrambling to patch, with some taking sites offline until fixes were in place. The bugs were addressed through a coordinated effort involving the PHP team, OpenAI, the research group Calif and Google-owned Wiz, whose CTO of Exposure Management and AppSec, Tomer Schwartz, took part in the response. Schwartz appeared on TechCrunch's Equity podcast with senior reporter Rebecca Bellan to describe the race from the inside and to argue that this episode is a preview rather than an ending.

Schwartz's central claim is that AI has changed the tempo of vulnerability discovery and exploitation. According to the researchers who found the bugs, the PHP incident is only the beginning. Wiz, he says, could not have found these flaws without AI: the company built custom tooling that ran through dozens of ideas over a single weekend. That detail is worth pausing on for anyone who has run security research teams. Testing dozens of hypotheses is ordinarily a weeks-long exercise in manual review and fuzzing setup; compressing it into a weekend implies the bottleneck has moved from generating candidate attack ideas to triaging and validating what the tooling surfaces.

One of the more striking aspects of the case is longevity. A bug sat in plain sight for more than a decade before conditions made it exploitable. This is a familiar pattern in widely deployed open source code: a latent defect is harmless under the assumptions of the era it was written in, then becomes dangerous when surrounding code, runtime defaults or attacker tooling shift. PHP's position as the substrate for a large share of the web, including the WordPress ecosystem, means flaws of this kind land on an enormous, heterogeneous install base, much of it run by people without dedicated security staff.

Disclosure logistics also featured in the conversation. Schwartz discussed why leaks are close to inevitable in coordinated disclosure and how Wiz planned for that possibility. Coordinated disclosure typically depends on embargoes shared across vendors, cloud providers and researchers, and the more parties in the room, the harder secrecy becomes. Hand patching, Schwartz argues, will not scale against this backdrop. The practical reading is that automated dependency updates, faster patch pipelines and better inventory of what is actually running in production matter more than ever, since an organization cannot patch software it does not know it has.

The episode also covers guidance for early-stage startups building with security in mind from the start, and addresses who remains unpatched after the PHP fixes, including whether small business owners running WordPress sites should be concerned. Those operational questions are likely the most immediately useful part for teams that do not have a Wiz-sized research group at their disposal.

It is worth separating what is established from what is projection. The existence of the two critical PHP bugs, their severity, the roughly 70% web share figure, the decade-plus latency of one flaw, the weekend tooling effort and the involvement of the PHP team, OpenAI, Calif and Wiz all come from the reporting around this episode. The broader 'Vulnpocalypse' framing is Schwartz's characterization of where things are heading, not a measured forecast. In practice, AI-assisted discovery cuts both ways: defenders gain faster hypothesis generation, and so do attackers, and the asymmetry often favors whoever can act on a finding first. A likely trade-off for vendors is that more frequent, smaller coordinated releases become the norm, trading the drama of emergency patches for a steady operational load on maintainers, many of whom are volunteers.

Why it matters: Any team shipping software that depends on widely used open source components, which is nearly every team, is exposed to a discovery cycle that AI tooling is accelerating. The concrete change for developers and product owners, as an inference from the discussion, is that patch velocity, software inventory and automated update pipelines shift from good hygiene to baseline requirements. Small operators on platforms like WordPress are the most exposed, since they carry the largest install base and the thinnest security staffing.

Read at TechCrunch · AI

Based on reporting from the original publisher. Visit the source for full context and later updates.

Publisher excerpt

The PHP scare may have faded from your feed, but the cleanup is far from over. Last week, two critical bugs in the open source language behind roughly 70% of websites left organizations scrambling to fix their systems, with some taking sites offline until the vulnerabilities could be patched. According to researchers who found the […]