Resources / AI
Problems in AI, and how we would solve them.
Problems we keep meeting in production AI: agents that stall, retrieval that invents citations, tool chains that leak. Each article names the problem, cites the evidence and proposes an architecture.
- AI agents3 min read
Why AI agent pilots stall before production
ProblemAgent pilots impress in demos and stall before production because nobody can show they work, limit what they can do, see what they did, or overrule them.ProposalDraw a hard decision boundary: the model scores and drafts, deterministic code decides and acts, every step is logged with its reasoning, and a person can overturn any decision in one click.Read the proposal - AI security3 min read
Prompt injection is an architecture problem
ProblemAny agent that reads untrusted content and can act can be steered by instructions hidden in that content, and no prompt or filter reliably separates data from commands.ProposalAssume injection will happen and limit what it can achieve: keep untrusted content away from privileged decisions, give tools least privilege, put consequential actions behind code-level checks, and break the path to exfiltration.Read the proposal - AI security3 min read
MCP servers are your new supply chain
ProblemAgents trust tool descriptions and outputs from MCP servers they install like packages, so a poisoned description, a compromised server or an over-privileged token can steer or rob them.ProposalPut an MCP gateway between agents and servers: pinned and reviewed tool manifests, per-user scoped tokens, sandboxed local servers, output sanitising, egress allowlists and a full audit log.Read the proposal - Retrieval3 min read
RAG still hallucinates. Grounding takes more than retrieval
ProblemRetrieval-augmented systems still invent facts, cite passages that do not support the claim, miss the decisive document and agree with false premises.ProposalTreat grounding as a pipeline with checks: structure-aware chunking, hybrid retrieval and reranking, citations at passage level that are verified after generation, abstention when evidence is missing, and evaluation of retrieval and answers separately.Read the proposal - Cost3 min read
Your AI bill is a routing problem
ProblemAI applications send every request to a frontier model with a long prompt, so cost and latency grow with usage while most requests never needed that capability.ProposalRoute each request to the cheapest model that passes your evaluation for that kind of request, cache shared prompt prefixes, batch what can wait, and distil a small model for the highest-volume task.Read the proposal - Models3 min read
Fine-tune, retrieve or prompt? A decision guide
ProblemTeams pick fine-tuning, RAG or custom training by reputation rather than by what the problem needs, and spend months on the wrong one.ProposalDecide by what must change: prompts for instructions, retrieval for knowledge that changes or must be cited, fine-tuning for behaviour, format and cost, and custom models only where the task or data is truly unusual; always prove it on an evaluation set.Read the proposal - Governance3 min read
The EU AI Act deadlines moved. Build the evidence trail anyway
ProblemTeams read the AI Act delay as a reprieve, while the evidence the Act requires, from risk management to logging and human oversight, takes longer to build than to document.ProposalGenerate compliance evidence as a by-product of engineering: versioned data and models, evaluation in CI, request-level logs, documented human oversight and incident handling, mapped once to the Act, NIST AI RMF and ISO/IEC 42001.Read the proposal - Media integrity3 min read
Proving a photo is real when anyone can edit one
ProblemGenerative editing makes convincing image manipulation cheap, and morphed ID photos can let two people pass as one, while detectors alone lose ground with every new model.ProposalCombine provenance and detection: fingerprint originals at capture or registration, anchor the fingerprint and its owner in a tamper-evident registry, attach Content Credentials, and run region-level tamper detection on anything without provenance.Read the proposal
One company at a time