Dane Knecht
Related people
Episodes
Notes
No Priors: Artificial Intelligence | Technology | Startups · Democratizing Security for the Agentic Web with Cloudflare CTO…
-
Should Cybersecurity AI Models Be Open Source?
Dane Knecht says cybersecurity AI has to be open over time, because the people who need to defend will not all sit inside a few labs.
The same model can attack and defend. Safety guardrails do not split those uses into a clean pair, so an open defensive model is also an offensive one. Keeping the capability inside a few closed frontier models fails the other way. Those labs cannot hand it to every person who has to defend a system. Knecht's view is that the tools spread anyway. That is how open source tends to work. Attackers will get them, and so will the biggest companies. A closed release leaves out everyone else who needs the defense.
-
Who Should Pay When AI Agents Read the Web?
Dane Knecht says the site owner should get paid when an AI agent reads the page, because agents and people are sharing a web that was priced for humans.
The web is in a messy middle. Agents try to parse pages built for people who look at ads and move slowly, and an agent can open a thousand pages in the time a person opens one. Humans and agents should not have to meet in the same place. Layers that translate a human page for an agent burn tokens, and APIs have only ever covered a small fraction of the internet. Agents want the rest, so they crawl the human web. That crawl has to get cheaper for the site, and the site has to be paid. A publisher such as a Substack already knows the CPM its
-
How Are AI Agents Finding Major Security Holes?
Dane Knecht says AI is now finding serious internet holes on a weekly rhythm, including a cryptographic key whose seed was the number four.
Log4j-scale incidents used to show up about once a year, and Cloudflare treated each one as all hands on deck for paying customers and free customers, because it sees that protection as a public good. Those incidents now happen weekly. One key that protects TLS and web PKI connections was created from the seed four. Cryptography needs randomness. Cloudflare keeps lava lamps so its own keys get real randomness. A key that starts from a single number can be reverse engineered, and a model does not sleep the way one red-team engineer does. Knecht says he t
-
Should I Put My Site Behind Cloudflare?
Dane Knecht says people who cannot afford four red teams, or who never had a red team and still run important systems, should put their sites and infrastructure behind Cloudflare and build on Cloudflare. When major vulnerabilities come out, like the React one last year, Cloudflare gives the protection to everyone, including customers beyond the paid plans with the advanced firewall and WAF capabilities. Cloudflare treats that protection as part of the utility, in the same way it wants to provide performance for everyone.
Before Cloudflare started, probably 3,000 sites in the world had a large-scale CDN to cache their stuff at the edge. Now hundreds of millions of sites do. Cloudflare has al
-
Why Is Skipping AI a Red Flag at Cloudflare?
Dane Knecht says skipping AI in your own life is a red flag when Cloudflare hires. Curiosity is one of the main things the company looks for, and drive, curiosity, and empathy matter even more for the next people it brings in. Someone that curious has little room to ignore AI. Fear and skepticism are the usual reasons people hold back, which Sarah Guo put to him as the reason the habit still has to be a principle at the company.
Cloudflare is inventing how the work gets done. The number of sites behind it, and the free services it gives away, make off-the-shelf technology impossible at that scale. New people have to come in and invent. Hiring for that starts with curiosity.
-
Why Was Cloudflare's Restructuring So Hard?
Dane Knecht says the restructuring was hard, and saying goodbye to some colleagues was the first time Cloudflare ever did it. Cloudflare had always been careful not to overhire. The reason was different: AI was transforming how some teams work, so Cloudflare thought of the operation like a factory floor that had to be rebuilt and rethought.
A product manager sending an email to customers used to involve three teams. They went to BI for a cohort of people, to the marketing team to design the email and get the voice right, and to the marketing operations team to get the email out. Redesigning the process took hard changes that forced people to rethink the right way to do it, and the new path
-
How Does Secure Software Start?
Dane Knecht says secure software starts by shrinking the attack surface: fewer places for a vulnerability, simpler software, and less code. Formal verification, safer programming languages, and safer runtimes will come later. Early AI is very verbose, likes to create complexity, and has not been helping Cloudflare build more secure software. He thinks it will get there. Human judgment still has to make sure the design can stand securely.
The work has to be solved at every level, the way humans did before agents. Product engineers need training. Security teams still need checks and balances and a second set of eyes. Cloudflare has been writing its principles down. A document called the Codex
-
How Long Will the AI Security Spending Surge Last?
Dane Knecht says the surge lasts about 18 months. Companies will pull security spending forward, on red teams that try to break systems and on defenses that get ready for those breaks. He does not expect that mix to last forever. After about 18 months, the same people go back to building, and new software productivity goes way up.
Security stays on everybody's mind through that stretch. Shared libraries and services have vulnerabilities that used to take real effort to find. Teams are also patching 70 years of software already in the world, so a lot of the work is catch-up while new software ships faster. The software development lifecycle and the agent development lifecycle have to change.