AI-assisted analysis. See our editorial policy.
Human editorial review not recorded
Ken Beer names the problem that makes key management genuinely hard, and it is not the cryptography: the key at the top of a hierarchy must be in plain text, all the time, for the system to remain available (4:46).
Everything below it can be encrypted by something above. The root cannot, or there is nothing to start from. That single constraint shapes every design decision in the field — hardware security modules exist to hold that plaintext key somewhere physically defensible, because the mathematical problem was already solved and the operational one was not.
Harvest now, decrypt later
The threat model that has moved from theoretical to planning-relevant gets a precise name here (11:25).
An adversary captures encrypted traffic today, stores it, and decrypts it later when the capability exists. The exposure is not to a future secret but to everything transmitted now, which means the decision about when to migrate cryptography was already made — data with a long confidentiality requirement is being harvested at whatever protection it has today.
The specific concern is asymmetric algorithms underpinning transport security and digital signatures (10:12). Symmetric encryption is comparatively resilient; the key exchange that establishes it is not.
This inverts the usual reasoning about upgrade timing. Ordinarily you migrate when the threat arrives. Here the threat's arrival is retroactive, so the deadline for anything that must stay confidential for a decade has already passed.
The point about delegation
The strategic argument is aimed at where teams should not be spending effort: rather than going through your source code to work out how to upgrade the cryptography, delegate it (23:08).
That is a vendor's argument and it happens to be right. Cryptographic agility — the ability to change algorithms without rewriting applications — is not something most organisations can retrofit, because the assumptions are embedded in code written by people who have left, in libraries nobody owns, in protocols chosen a decade ago.
Where it sits behind a managed interface, the migration is somebody else's engineering problem. Where it sits in application code, it is an audit of everything you have ever shipped.
The agentic connection
One remark ties this session to the rest of the conference: certificate management is being reshaped by the explosion of agentic AI (6:54).
The logic is straightforward and its consequences are not. Certificates express the identity of a server or client (18:11). When the things establishing connections were servers, their number was bounded by what you deployed. When agents establish connections, the population is bounded by nothing in particular, and each one needs an identity that can be issued, rotated and revoked.
Certificate infrastructure was built for a world with a countable number of endpoints. That assumption is now the constraint, and this is one of the few places anyone said so out loud.
Talk chapters
Key takeaways
- 01
The root of a key hierarchy must remain in plain text for availability, which is why hardware security modules exist — the mathematics was solved, the operational problem was not. 4:46
- 02
Harvest now, decrypt later means the exposure is to traffic transmitted today, so the migration deadline for long-confidentiality data has already passed. 11:25
- 03
The specific vulnerability is in asymmetric algorithms underpinning transport security and signatures rather than in symmetric encryption. 10:12
- 04
His strategic argument is delegation: auditing your own source code for cryptographic assumptions is a task most organisations cannot complete. 23:08
- 05
Certificate management is being reshaped by agentic AI, because certificate infrastructure assumed a countable number of endpoints. 6:54
Entities mentioned
Organizations
Related talks
The practical counterpart to the argument made elsewhere this season that specification is what contains model entropy. Raval and Harris name the failure they are addressing precisely — a prompt-and-pray loop in which working code arrives with no record of what the model assumed, which requirements were fuzzy, what design was chosen or why, leaving nothing to review and nothing to iterate against when a defect surfaces months later. Their answer is three committed markdown artefacts: requirements written in a structured requirements syntax with acceptance criteria attached to each user story, a design document carrying technical decisions together with the reasoning behind them, and a task list whose entries cite the requirement numbers they satisfy. The traceability is the point — a reviewer questioning a decision in a pull request can follow it back through the task to the design to the requirement, all in the same repository. Notably they keep the human between each phase rather than after it, with the agent surfacing ambiguity as questions before proceeding.
Drawn from a year of engagements with more than a hundred companies, this is the most direct challenge in the season's programme to the assumption that faster code generation produces faster delivery. Mishra and Raja open with external evidence rather than their own: an industry study putting realised velocity gains in the ten to fifteen per cent range, and a controlled experiment in which developers using AI estimated themselves roughly a fifth more productive while measurement showed them a fifth slower. Their diagnosis is that both prevailing working styles fail for opposite reasons. Handing an ambiguous problem to an agent and awaiting a finished result produces a volume of code the developer must nonetheless sign for and cannot confidently review, so it stalls before production. The senior engineer's alternative — decomposing the work personally and inserting AI into narrow slots — keeps the intellectual load exactly where it was, and leaves the surrounding process untouched, so hours saved in editing are consumed by the meetings that process still requires.
Brooker builds the definition from the bottom up rather than asserting it, using a deliberately absurd arithmetic task to separate three categories: what a model computes reliably as a fixed function of its input, what merely needs to arrive in the system prompt, and what genuinely requires reaching into the world. Only the third category justifies a tool, and the distinction matters because most production disappointment comes from tools built for the first two. His working definition follows — a system given a goal that loops between inference and tool calls until it reaches one — with the observation that modern agents increasingly embed code in their definitions, not for expressiveness but because replacing inference steps with deterministic code improves reliability while lowering both latency and cost. The remainder covers what production actually demands around that loop: somewhere to run, memory that persists preferences, a gateway to internal and external tools, evaluation, and formal methods applied to policy.
The most concrete attempt this conference season to answer a question the agentic coding sessions mostly leave open: if commit counts and hours saved are the wrong measures, what replaces them? Otto's account is unusually specific about why the obvious alternative fails — summing the small time savings a platform team delivers produces figures exceeding a hundred per cent of a developer's time, and a minute returned is not code in production. Their replacement borrows from Amazon's retail supply chain, where cost to serve measures what it takes to place a package on a doorstep, and applies the same shape to software: total cost divided by units of delivery, with the unit chosen to fit the team. The supporting research is the more quotable finding — across tens of thousands of developers over five years, individual velocity reverts to the team's mean, making team velocity the strongest predictor of both individual output and perceived productivity, which is the empirical case against measuring individuals at all.
The framing statistic is organisational rather than technical: around eighty per cent of organisations expected to have platform engineering teams going into 2026, up from about forty-five per cent a couple of years earlier. The interesting part is the doubling. The problem described is teams solving the same problems separately, producing inconsistency and redundancy — dangerous not because of duplicated effort but because each independent solution has its own security properties, and the organisation's real posture is the weakest rather than the average. The most valuable content is that two financial services organisations went in diametrically opposite directions on workload identity and both are described as working, which implies the choice is determined by context rather than by a general answer. The honest note follows immediately: even with standardised patterns the result remains fragmented.
The line that explains this session comes from the customer in the last ten minutes: they are preparing for a world where metadata is how agent-based systems find the data they need and access it through the controls being built. That relocates a function — governance has spent two decades as compliance activity describing data that people locate by other means, and if agents navigate by the catalogue then the catalogue stops describing the access path and becomes it. An incomplete catalogue is a documentation problem when humans can ask a colleague; an agent has no such workaround. The most honest moment addresses the perennial failure that rules get written and ignored, with enforcement rather than publication as the argument. Generated descriptions and greyed-out classification suggestions divide the labour correctly, keeping a person accountable while removing the burden of finding candidates.