Chain of Thought | AI Agents, Infrastructure & Engineering

Conor Bronsdon

AI is reshaping infrastructure, strategy, and entire industries. Chain of Thought is the podcast where builders reason through what's changing. Host Conor Bronsdon sits down with the engineers and founders shipping AI in production to get past the hype into what's working and what isn't. Episodes cover model infrastructure, inference, agent frameworks, evaluation, and developer tools. Guests have come from NVIDIA, Google DeepMind, AMD, Databricks, Vercel, and more. Every episode carries a full transcript and show notes at chainofthought.show. New episodes weekly. Conor Bronsdon is an independent consultant and angel investor in AI infrastructure and developer tools. He led technical ecosystem at Modular, acquired by Qualcomm in 2026; led developer awareness at Galileo, acquired by Cisco; and ran developer marketing at LinearB, where he was GM of the Dev Interrupted podcast and community. Views expressed by the host and guests are their own.

  1. 19h ago

    Writing Code Was Never the Bottleneck | CloudBees CEO Moritz Plassnig

    Enterprises are spending tens of millions, sometimes hundreds of millions, on tokens for their engineers. Ask what they're shipping as a result and, Moritz Plassnig says, the answer can be just a few more pull requests. In big companies, he says, writing code was never the bottleneck. Builds run for hours. Policies still require a human to review every line. Releases wait weeks for security and governance sign-off. Generating more code doesn't move any of that. Moritz is the CEO of CloudBees and the founder of Codeship, the CI/CD platform CloudBees acquired in 2018. He tells Conor where enterprise AI adoption stalls, why the people inside a company take longer to adapt than the technology, and what has to be rebuilt once there's a hundred times more code. We cover:- Why enterprise AI adoption is moving faster than anything Moritz has seen, and still slower than it looks from San Francisco and X- The large bank that rolled out several agentic coding tools while its policy still says every line of code must be reviewed by a human- Measuring tokens against pull requests, then production, then revenue, and the thought experiment of giving a company ten times the engineers- Why Moritz tells companies not to over-focus on ROI at first, and what relearning how to swim taught him about that- Policies as code, and why developers should keep their choice of models and tools- Why enterprises are signing with several model vendors at once, and why the moment the US government moved to regulate who gets the best models still shapes those deals- GitHub losing its place as the default, and why CI has to be rethought for a hundred times more code: predictive test selection and not rerunning tests that already passed- Rethinking pull requests, feature flags and the rest of software delivery from first principles Connect with Moritz Plassnig:Twitter/X: https://x.com/moritzplassnigLinkedIn: https://www.linkedin.com/in/moritzplassnig/CloudBees: https://www.cloudbees.comCloudBees on X: https://x.com/CloudBees Connect with Chain of Thought host Conor Bronsdon:Newsletter: https://newsletter.chainofthought.show/Twitter/X: https://x.com/ConorBronsdonLinkedIn: https://www.linkedin.com/in/conorbronsdon/YouTube: https://www.youtube.com/@ConorBronsdon More episodes: https://chainofthought.show Chapters:(0:00) AI code volume will force a first-principles rethink(0:49) Enterprises are all in on AI-written code(3:26) Real adoption speed versus the X hype(13:39) Companies handling the AI pivot well(17:12) Code exists to deliver value(23:19) Shipping more code versus business outcomes(28:37) Different advice for startups and enterprises(34:42) Transforming legacy pipelines without shutting them down(38:25) Staying flexible across multiple model vendors(43:28) GitHub outages and how CI is changing(49:41) The vestigial steering wheel of software(53:35) Rethinking everything as intelligence gets cheaper Thanks to Inngest, presenting sponsor of season four of Chain of Thought. Agents in production run long - they call models and wait on APIs and people. But the longer agents run, the more they break. Inngest handles that with durable execution. You build your agent as steps in TypeScript, Python, or Go. When a step fails, Inngest retries it with exponential backoff, and completed steps are saved and skipped. Try it out: https://inngest.link/cot-pod Thanks to CoralBricks for sponsoring this episode. CoralBricks serves open models through one API, fast and at low cost: DeepSeek V4.1 Flash streams at over 300 tokens a second. Get $5 in credits on signup, plus another $5 on your first purchase: https://www.coralbricks.ai/r/Z6V2RE

    Writing Code Was Never the Bottleneck | CloudBees CEO Moritz Plassnig
  2. 2d ago

    This is the worst AI tutor we'll ever have | DataCamp

    DataCamp tested an early version of its AI tutor against its classic courses, and the tutor won: willingness to pay, engagement and satisfaction were all higher. It was also easily ten to twenty times more expensive per hour of learning than DataCamp's previous setup. Financially, Jonathan says, shipping the better product would have made the business worse. He expects that cost to fall until it looks like the cost of serving DataCamp's interactive courses. The only question, he says, is whether that takes four to eight weeks or another year. He's building the tutor on that bet. Jonathan Cornelissen is co-founder and CEO of DataCamp and is joined by Yusuf Saber, DataCamp's Chief AI Officer and the founder of Optima, the learning company DataCamp acquired. They walk Conor through how the tutor works: teaching behaviors it has to show on cue, evals that test each one repeatedly, a mix of models picked for speed and cost, and a latency budget measured in milliseconds. We cover:- How DataCamp turns what great human teachers do into about 30 core behaviors, and how the tutor gives feedback: one main issue at a time, hints phrased as questions, and asking before it reveals a solution- The A/B test against DataCamp Classic, and why Jonathan calls today's tutor the worst one DataCamp will ever have- The models behind the tutor: Claude Sonnet handles about 70% of the work, Gemini Flash used to handle much of the fast work, and a cheaper tier is starting to take on more- Why open models keep failing on speed, hosting and caching, even after one outperformed DataCamp's best model at the time- Going from almost nothing to a run rate of over a million hours of learning a year, and why the same three-hour course can take one learner half an hour and another six hours- Pre-generating responses while the learner thinks, and why a stronger model lets them aim for cache hit rates in the mid-90s- What Bloom's tutoring research says, why doing better on a test isn't proof of learning, and what happens to junior roles when AI takes the grunt work Connect with Jonathan Cornelissen:LinkedIn: https://www.linkedin.com/in/jonathan-cornelissen/Twitter/X: https://x.com/CornelissenJoDataCamp: https://www.datacamp.comDataCamp on X: https://x.com/datacamp Connect with Yusuf Saber:LinkedIn: https://www.linkedin.com/in/yusuf-saber-ba380b15/ Connect with Chain of Thought host Conor Bronsdon:Newsletter: https://newsletter.chainofthought.show/Twitter/X: https://x.com/ConorBronsdonLinkedIn: https://www.linkedin.com/in/conorbronsdon/YouTube: https://www.youtube.com/@ConorBronsdon More episodes: https://chainofthought.show Chapters:(0:00) Better product, worse economics(0:59) How the tutor helps a struggling learner(7:09) From static courses to AI tutoring(9:25) Multi-agent architecture behind the tutor(13:10) Scaling costs and testing open models(18:30) Economics of cost per learning hour(22:13) What great looks like for DataCamp(25:33) Latency budgets and precomputed lesson plans(32:53) Measuring learning outcomes with Bloom's research(39:20) AI skills gap for early-career workers(43:22) Evals built on tutor behavior expectations(50:12) Closing thoughts on how skills are changing Thanks to CoralBricks for sponsoring this episode. CoralBricks serves open models through one API, fast and at low cost: DeepSeek V4.1 Flash streams at over 300 tokens a second. Get $5 in credits on signup, plus another $5 on your first purchase: https://www.coralbricks.ai/r/Z6V2RE Thanks to Walrus, presenting sponsor of season four of Chain of Thought. Walrus Memory gives AI agents portable, verifiable memory that carries context across apps, sessions, and other agents. Get started at https://walrus.xyz/cot Thanks to Inngest, presenting sponsor of season four of Chain of Thought. Agents in production run long - they call models and wait on APIs and people. But the longer agents run, the more they break. Inngest handles that with durable execution. You build your agent as steps in TypeScript, Python, or Go. When a step fails, Inngest retries it with exponential backoff, and completed steps are saved and skipped. Try it out: https://inngest.link/cot-pod Thanks to WisdomAI for sponsoring this episode. In WisdomAI's latest benchmark, Claude Code alone topped out at 73% accuracy; WisdomAI closed the gap, delivering accurate answers 3x faster at about a third of the cost. Get "The Blueprint for Trusted AI Analytics": https://wisdom.ai/benchmark

    This is the worst AI tutor we'll ever have | DataCamp
  3. 4d ago

    The Infrastructure Behind Agents That Run for Months

    Tony Holdstock-Brown is the CEO and co-founder of Inngest, which built durable workflows for background jobs before the agent wave arrived. He says most Inngest customers no longer log into the UI. Their agents check on what they have built through the API, the CLI and MCP, and people who build with agents activate at about 93 percent, against roughly 40 percent before. Holdstock-Brown defines a long-running agent as one that lives for months and builds up its own context. Running one means orchestrating sub-agents, sandboxes and state for every customer, and he argues that orchestration layer becomes the product. He and Conor Bronsdon also get into why agent load compounds, how to score an agent from production events, and what Inngest learned from running out of compute. If your users never open your product themselves, how it behaves for their agents is the whole experience. We cover: Why durable workflows, a series of steps with checkpoints, mapped cleanly onto agentsHoldstock-Brown's definition of a long-running agent: it lives for months and builds its own contextInngest's shift from UI traffic to API, CLI and MCP, and the 93 percent activation rate among people building with agentsHow agent load compounds when every user's agent makes thousands of API requests a dayTracking events and outcomes for every agent run, a lesson Holdstock-Brown ties to KISSmetrics and A/B testingExplicit step APIs compared with compiler directives, and why a step is close to tracingRunning out of compute: bare metal, bandwidth limits and building on FoundationDBNew episodes, the ideas behind them and Conor's essays land in the Chain of Thought newsletter first. Subscribe: https://newsletter.chainofthought.show/ Chapters:(0:00) Why agents need durable infrastructure(1:13) How Inngest started before agents(3:53) Defining a long-running agent(6:08) Orchestration becomes the product(8:40) Customers' agents replace the UI(10:40) Reliability when agent load compounds(16:40) State, memory and personal agents(20:50) Personal agents for mainstream users(28:07) Step APIs vs compiler directives(31:27) Code review agents and security(40:38) Running out of compute(46:55) Building your first long-running agent Links from the episode: How the Claude Code team does code review (Anthropic): https://claude.dev/blog/solving-code-review-at-anthropic/Claude Duet announcement: https://claude.com/resources/articles/claude-duetClean Code by Robert C. Martin: https://www.informit.com/store/clean-code-a-handbook-of-agile-software-craftsmanship-9780132350884Software Engineering is Dead. Long Live Product Engineering (Conor's essay): https://newsletter.chainofthought.show/p/software-engineering-is-dead-longOpenClaw: https://openclaw.aiChain of Thought with Claire Vo: How AI Velocity is Rewriting the Rules for Engineering Leaders: https://share.transistor.fm/s/5178eb0aChain of Thought with Jaime DeLanghe: Slack Wants to Be the Context Harness for Code: https://share.transistor.fm/s/274962beChain of Thought with Laurie Voss: AI Codes: Product Engineers Decide What to Build: https://share.transistor.fm/s/6deea963Chain of Thought with Wen Sang: Genspark's Bet: AI Agents Become the Users of Software: https://share.transistor.fm/s/c898a0e8Connect with Tony Holdstock-Brown: X: https://x.com/itstonyhbInngest: https://inngest.link/cot-podConnect with Chain of Thought host Conor Bronsdon: Newsletter: https://newsletter.chainofthought.show/Twitter/X: https://x.com/ConorBronsdonLinkedIn: https://www.linkedin.com/in/conorbronsdon/YouTube: https://www.youtube.com/@ConorBronsdonMore episodes: https://chainofthought.show Thanks to Inngest, presenting sponsor of season four of Chain of Thought. Agents in production run long - they call models and wait on APIs and people. But the longer agents run, the more they break. Inngest handles that with durable execution. You build your agent as steps in TypeScript, Python, or Go. When a step fails, Inngest retries it with exponential backoff, and completed steps are saved and skipped. Try it out: https://inngest.link/cot-pod

    The Infrastructure Behind Agents That Run for Months
  4. Nov 8

    Asking Your Team to Automate Their Jobs With AI | Cloudera's Shayde Christian

    Shayde Christian's team at Cloudera has built AI agents that can open a sales opportunity in the CRM and send a customer a 30% discount on training. He still isn't authorized to switch them on in production. Cloudera runs as conservatively as the regulated customers it serves, and the business owners want more proof of reliability and accuracy first. The AI tool his team does lean on every day is, in his words, barely AI. Full transcript and chapters: https://chainofthought.show/podcast/82-asking-your-team-to-automate-their-jobs-with-ai-clouderas-shayde-christian/ Shayde Christian is Chief Data and Analytics Officer at Cloudera. He walks Conor through how support-case data became a signal for which customers need training, how Cloudera uses a public model to vectorize questions and private agents to execute or generate SQL, and why he developed new career paths before asking his data team to automate their own jobs. We cover:- How three non-break-fix support cases a quarter on a product like NiFi flag a customer who needs training, and Shayde's Bellagio fountain idea of dashboards that disappear once a decision is automated- Getting customer success managers up to speed on a new account in an hour or two instead of a week, and covering eight or nine accounts instead of two to four- How the query tool uses a public model to vectorize the question and agents to run existing SQL and Python; when no match is found, the agent can generate SQL, with a disclosure and feedback check- 200 power users for $2,000 a month, and why Shayde's advice is to ask whether you need AI, not just where to run it- Why data governance is still hard, and the table renaming, synonym columns and metadata work it took for AI to understand business meaning- Cloudera's new Mistral partnership: why Shayde says "Mistral reigns" in Europe and the Middle East, the sovereign AI stack with NVIDIA and Dell, and a proof of concept for agents that act across a company's SaaS systems- Three career paths for a data team whose work is being automated: AI engineer, "parachutist" and "insight assassin"- The human skills that matter more as AI takes routine work, and his advice to new graduates: automate your last job Connect with Shayde Christian:LinkedIn: https://www.linkedin.com/in/shaydechristianCloudera: https://www.cloudera.com Links from the episode:Cloudera and Mistral partnership announcement: https://mistral.ai/news/mistral-x-cloudera/MotionBench: https://motionbench.sh Connect with Chain of Thought host Conor Bronsdon:Newsletter: https://newsletter.chainofthought.show/Twitter/X: https://x.com/ConorBronsdonLinkedIn: https://www.linkedin.com/in/conorbronsdon/YouTube: https://www.youtube.com/@ConorBronsdon More episodes: https://chainofthought.show Chapters:(0:00) Why automating data work makes the team more valuable(1:05) Actionable signals buried in support cases(5:03) Data privacy culture as AI blocker(6:04) A shared customer view instead of more dashboards(13:08) Three levels of AI business value(15:47) Keeping data private with frontier models(23:55) AI coding tools and data masking(29:05) Evaluating accuracy before agentic actions(32:46) Why Mistral leads in Europe(37:10) Asking the team to automate their jobs(44:24) Change management starts with mindset(48:26) Career advice for new graduates Thanks to Walrus, presenting sponsor of season four of Chain of Thought. Walrus Memory gives AI agents portable, verifiable memory that carries context across apps, sessions, and other agents. Get started at https://walrus.xyz/cot Thanks to CoralBricks for sponsoring this episode. CoralBricks serves open models through one API, fast and at low cost: DeepSeek V4.1 Flash streams at over 300 tokens a second. Get $5 in credits on signup, plus another $5 on your first purchase: https://www.coralbricks.ai/r/Z6V2RE Thanks to Inngest, presenting sponsor of season four of Chain of Thought. Agents in production run long: they call models and wait on APIs and people. But the longer agents run, the more they break. Inngest handles that with durable execution. You build your agent as steps in TypeScript, Python, or Go. When a step fails, Inngest retries it with exponential backoff, and completed steps are saved and skipped. Try it out: https://inngest.link/cot-pod Thanks to WisdomAI for sponsoring this episode. In WisdomAI's latest benchmark, Claude Code alone topped out at 73% accuracy; WisdomAI closed the gap, delivering accurate answers 3x faster at about a third of the cost. Get "The Blueprint for Trusted AI Analytics": https://wisdom.ai/benchmark

    Asking Your Team to Automate Their Jobs With AI | Cloudera's Shayde Christian
  5. Nov 1

    Agents Are Overtaking Humans in Search: Why Search Must Be Rebuilt

    Saahil Jain joined You.com as a founding engineer in 2020, when it was a consumer search engine competing with Google. Now its CTO, he builds search for AI agents, and he argues that however smart models get, they will still need tools to find out what is happening in the world. His analogy: we built the self-driving cars before the roads. He walks through which work he expects to move inside the model, why better models made harnesses matter less, and why cheaper open models and routing are giving harnesses a new job. He says he changed his mind on that last point in the past few months. Then we get into how agents search differently than people do: thousands of queries at once, recall over ranking, and a much wider range of latency needs. Saahil traces a query from fan-out to the index to re-ranking, and explains why You.com judges search by whether the agent answered the question, and why he trusts customer evals and benchmarks that change over time more than static public ones. We cover: - Core intelligence versus augmented intelligence, and the self-driving cars that arrived before the roads- Which tasks Saahil expects models to absorb, and where unverifiable rewards slow them down- Why frontier labs keep going vertical while open-model labs give weights away- How stronger models shrank the gap between good and bad harnesses, and how routing across open models can revive them- How You.com post-trains models with its search tool while trying not to overfit to any one model- How agent search differs from human search: query volume, context, ranking and latency- The life of a query: subqueries, lexical and semantic indexes, re-ranking and enrichment- Why static benchmarks can be memorized, and what dynamic benchmarks and customer evals measure instead Mentioned in this episode:- I Paid an AI Agent $8 to Write About its ‘Life’: https://newsletter.chainofthought.show/p/i-paid-an-ai-agent-8-to-write-about- MotionBench: https://motionbench.sh Connect with Saahil Jain: - LinkedIn: https://www.linkedin.com/in/saahiljain/- Twitter/X: https://x.com/saahil9jain- You.com: https://you.com- You.com API docs: https://you.com/docs/quickstart Connect with Chain of Thought host Conor Bronsdon:Newsletter: https://newsletter.chainofthought.show/Twitter/X: https://x.com/ConorBronsdonLinkedIn: https://www.linkedin.com/in/conorbronsdon/YouTube: https://www.youtube.com/@ConorBronsdon More episodes: https://chainofthought.show Chapters:(0:00) Agents run a thousand searches at once(0:25) What survives as AI models get smarter(1:47) Is this the singularity?(4:26) Core intelligence versus augmented intelligence(6:33) Self-driving cars before the roads(7:27) What belongs inside the model(11:00) Where verifiable rewards run out(12:16) Can agents form relationships?(16:58) What model labs own, and what independents can build(20:21) What a harness is(22:39) The model is eating the harness(23:14) Open models and routing revive the harness(25:32) Open versus closed: what the labs keep(31:30) OpenAI, Baseten and the enterprise model budget(33:53) Co-evolving a search tool with the model labs(37:41) How agent search differs from human search(42:42) Life of a query, from fan-out to index(47:22) Measuring retrieval by whether the agent answered(50:21) Static benchmarks, freshness and customer evals(54:19) Why he changed his mind on harnesses Thanks to Walrus, presenting sponsor of season four of Chain of Thought. Walrus Memory gives AI agents portable, verifiable memory that carries context across apps, sessions, and other agents. Get started at https://walrus.xyz/cot Thanks to CoralBricks for sponsoring this episode. CoralBricks serves open models through one API, fast and at low cost: DeepSeek V4.1 Flash streams at over 300 tokens a second. Get $5 in credits on signup, plus another $5 on your first purchase: https://www.coralbricks.ai/r/Z6V2RE Thanks to Inngest, presenting sponsor of season four of Chain of Thought. Agents in production run long: they call models and wait on APIs and people. But the longer agents run, the more they break. Inngest handles that with durable execution. You build your agent as steps in TypeScript, Python, or Go. When a step fails, Inngest retries it with exponential backoff, and completed steps are saved and skipped. Try it out: https://inngest.link/cot-pod Thanks to WisdomAI for sponsoring this episode. In WisdomAI's latest benchmark, Claude Code alone topped out at 73% accuracy; WisdomAI closed the gap, delivering accurate answers 3x faster at about a third of the cost. Get "The Blueprint for Trusted AI Analytics": https://wisdom.ai/benchmark

    Agents Are Overtaking Humans in Search: Why Search Must Be Rebuilt
  6. Oct 29

    How Cisco Secures Personal AI Assistants for 90,000 Employees | DJ Sampath

    Cisco is giving each of its 90,000 employees a personal AI agent. Some of those same employees are already connecting agents from outside the security boundary without asking IT. DJ Sampath, SVP and GM of AI Software and Platform at Cisco, walks through what had to be in place before the rollout: scanning every prompt, sandboxing each agent, keeping agents multi-tenant, scanning the skills and tools people bring in, and settling which identities an agent may act as. New episodes, the ideas behind them and Conor's essays land in the Chain of Thought newsletter first. Subscribe: https://newsletter.chainofthought.show/ We recorded at Cisco's San Francisco office, a block or two from OpenAI, where Cisco was an early Codex design partner. DJ explains how engineers rework their skills files with each model generation, how per-engineer PR bots review agent-written code, and why defenders now have to move at agent speed when an exploit can show up within minutes of a CVE going public. We cover: What Cisco's IT team required before rolling out MyAgent to 90,000+ employeesPrompt injection by email, and why the first integration everyone makes is the riskiestCisco's AI security taxonomy, DefenseClaw, and Project Quarry, a new open source language for describing agent behavior and policyHow the agents are actually used: daily briefs, meeting actions, live dashboards, 2 trillion tokens in a single weekWhy skills files need refactoring every model generationDiscovering and governing shadow agents with zero trust accessMulti-model code review, spec drift, and evals built from pull request tracesHow Cisco Cloud Control was built in about eight months with agentic codingChapters:(0:00) Guardrails tested inside Fortune 100 companies(0:21) Securing 90,000 employee AI agents(8:04) Cisco's open source agent security stack(12:01) Project Quarry: a policy language for agents(17:07) Personal agent use cases at Cisco(19:08) Early design partner on Codex(23:15) Reassessing workflows each model generation(27:27) Shadow agents and zero trust access(30:29) Security as an adoption accelerator(31:39) Multi-model code review for dependability(36:09) Defenders moving at agent speed(41:04) How Cisco Cloud Control was built Connect with DJ Sampath: X: https://x.com/djsampathLinkedIn: https://www.linkedin.com/in/djsampath/DJ on Cisco Blogs: https://blogs.cisco.com/author/djsamProject Quarry: https://blogs.cisco.com/ai/project-quarry-agent-security-detectionsCisco: https://www.cisco.com/Related episode: Cisco President and CPO Jeetu Patel, "Parenting Your AI Agents for Best Results": https://share.transistor.fm/s/43e97aa6 Connect with Chain of Thought host Conor Bronsdon: Newsletter: https://newsletter.chainofthought.show/Twitter/X: https://x.com/ConorBronsdonLinkedIn: https://www.linkedin.com/in/conorbronsdon/YouTube: https://www.youtube.com/@ConorBronsdonMore episodes: https://chainofthought.show Thanks to Inngest, presenting sponsor of season four of Chain of Thought. Agents in production run long: they call models and wait on APIs and people. But the longer agents run, the more they break. Inngest handles that with durable execution. You build your agent as steps in TypeScript, Python, or Go. When a step fails, Inngest retries it with exponential backoff, and completed steps are saved and skipped. Try it out: https://inngest.link/cot-pod Thanks to Walrus, presenting sponsor of season four of Chain of Thought. Walrus Memory gives AI agents portable, verifiable memory that carries context across apps, sessions, and other agents. Get started at https://walrus.xyz/cot Thanks to WisdomAI for sponsoring this episode. In WisdomAI's latest benchmark, Claude Code alone topped out at 73% accuracy; WisdomAI closed the gap, delivering accurate answers 3x faster at about a third of the cost. Get "The Blueprint for Trusted AI Analytics": https://wisdom.ai/benchmark Thanks to Svix, presenting sponsor of season four of Chain of Thought. Svix delivers billions of reliable webhooks for startups and the Fortune 500. Get started at https://link.svix.com/cot. Qualified startups get $12,000 in credits, and YC companies get $50,000.

    How Cisco Secures Personal AI Assistants for 90,000 Employees | DJ Sampath
  7. Oct 25

    For AI Agent Fleets, the Model Is Rarely the Bottleneck | Theory Ventures' Andy Triedman

    Andy Triedman is a partner at Theory Ventures, and he is building his own software factory, while assessing how startups are leveraging theirs. His read after running agent fleets: the models are frankly good enough for most of the work. The vast majority of things you'd ask an agent to do are not intelligence limited. They are context limited. We have infrastructure work to do - but we can be more ambitious! Full transcript and chapters: https://chainofthought.show/podcast/79-for-ai-agent-fleets-the-model-is-rarely-the-bottleneck-theory-ventures-andy-triedman/ So the hard part moves upstream. Deciding what the agents should build, what they need to know, and when they should stop and ask a person is now most of the job. We cover: Moving from solving one problem at a time to building systems that solve whole classes of problems, like generating and testing integrationsWhich tasks still need human approval, and which can be fully handed offWhy a short feature spec leaves too much unsaid, and how better questions up front cut review interruptionsKeeping shared context and company decisions current, because an agent takes everything it reads as trueInference ROI, and when well-scoped tasks can go to cheaper modelsSandboxes and simulated customer environments for agentsFirst steps for teams: shared skills, cloud agents and approval gatesProtecting customer data, and organizing agent memory from session traces to company prioritiesNew episodes, the ideas behind them and Conor's essays land in the Chain of Thought newsletter first. Subscribe: https://newsletter.chainofthought.show/ Chapters:(0:00) Agents are context limited, not intelligence limited(0:34) From coding agents to software factories(5:53) What still needs human approval(10:16) Agent loops that improve the product on their own(14:52) Where agent automation breaks down(17:59) The bitter lesson and engineering as experimentation(20:08) From writing code to directing agent fleets(24:18) Shared context and stale decisions(27:25) Inference ROI and cheaper models(30:23) Moving human time from code to customer intent(34:42) Sandboxes and simulated customer environments(37:26) First steps: shared skills and cloud agents(40:14) Protecting customer data and isolating agents(42:32) Context limits and hierarchical agent memory(47:05) What Theory Ventures looks for in founders Connect with Andy Triedman: LinkedIn: https://www.linkedin.com/in/andy-triedman/Andy at Theory Ventures: https://theoryvc.com/team-members/andy-triedmanTheory Ventures: https://theoryvc.com/Connect with Chain of Thought host Conor Bronsdon: Newsletter: https://newsletter.chainofthought.show/Twitter/X: https://x.com/ConorBronsdonLinkedIn: https://www.linkedin.com/in/conorbronsdon/YouTube: https://www.youtube.com/@ConorBronsdonMore episodes: https://chainofthought.show Thanks to Svix, presenting sponsor of season four of Chain of Thought. Svix delivers billions of reliable webhooks for startups and the Fortune 500. Get started at https://link.svix.com/cot. Qualified startups get $12,000 in credits, and YC companies get $50,000. Thanks to Walrus, presenting sponsor of season four of Chain of Thought. Walrus Memory gives AI agents portable, verifiable memory that carries context across apps, sessions, and other agents. Get started at https://walrus.xyz/cot Thanks to Inngest, presenting sponsor of season four of Chain of Thought. Agents in production run long - they call models and wait on APIs and people. But the longer agents run, the more they break. Inngest handles that with durable execution. You build your agent as steps in TypeScript, Python, or Go. When a step fails, Inngest retries it with exponential backoff, and completed steps are saved and skipped. Try it out: https://inngest.link/cot-pod Thanks to G2i for sponsoring this episode - for over a decade, they vetted and placed engineers at other companies, from startups to FAANG. Two years ago, they turned that same judgment inward, building their own bench to review RL environments, evals, and training data that models are trained on. Get access: https://fandf.co/3SFxVm6

    For AI Agent Fleets, the Model Is Rarely the Bottleneck | Theory Ventures' Andy Triedman
  8. Oct 14

    What Breaks When Agents Become the Customer | Svix CEO Tom Hacohen

    What breaks when agents become the customer? Most of the integration layer, APIs, SDKs and webhooks alike, was built for human developers and for servers that are always listening. Agents are neither. Full transcript and chapters: https://chainofthought.show/podcast/78-what-breaks-when-agents-become-the-customer-svix-ceo-tom-hacohen/ When an agent runs on a laptop, closing the lid takes down the tunnel its webhooks come through. When the lid opens, it gets a storm of retries from 12 hours earlier. Svix founder and CEO Tom Hacohen's answer is to let the agent poll for what it missed. Svix delivers webhooks for companies like Clerk and Resend, and it created Standard Webhooks, the spec OpenAI, Anthropic and Google Gemini have adopted. In this episode of Chain of Thought, Tom joins Conor Bronsdon to talk through designing APIs for agents that adapt, whether agents still need SDKs, and why he wants less code even when agents write it. He also introduces Poke, Svix's new spec for making event-driven APIs work for agents. This episode is sponsored by Svix. We cover: Why events are harder for agents than API calls: the sender controls them, and they arrive outside the agent's sessionHow polling endpoints and Svix's plugins for Hermes and OpenClaw let an agent pick up where it left offWhy Tom calls JSON an abomination and webhooks duct tape, and why worse is better anywayWhy Svix created Standard Webhooks when everyone implemented webhooks differently, and how the spec let Kong and ngrok add webhook verificationWhether agents still need SDKs, and why Svix's skills work like a solutions engineer while its docs stay written for humansWhy Tom wants less code even when agents write it, and what piling patches on patches costsPoke, Svix's early spec for letting agents subscribe to and trigger events across services, released so feedback can shape it while it's still malleableChapters:(0:00) Duct tape, retry storms and the case for simplicity(1:19) Why integrations got easier and harder at once(2:59) How EteSync's webhook requests led to Svix(5:22) Worse is better: why webhooks persist(7:15) What changes when an agent receives the event(9:07) What happens when the laptop closes(11:08) Hermes, OpenClaw and agents that poll(13:03) Letting agents choose their own subscriptions(14:42) The long tail of the web without APIs(18:38) Why Svix created Standard Webhooks(24:17) How Standard Webhooks is governed(26:16) Following real usage, including MCP(28:17) Designing APIs for agents that adapt(30:38) Do agents still need SDKs?(31:25) Skills as documentation for agents(34:52) Less code, even when agents write it(36:19) Agentic development and the offshoring analogy(37:34) Software factories and the plumbing underneath(40:57) Keeping up while the standards change(44:18) What Svix is launching with Poke(46:39) Why Svix released the Poke spec early(47:58) Building for your customers' customers(49:38) Clerk, Resend and webhooks in the background(52:22) Where to follow Tom Links from the episode: Svix: https://link.svix.com/cotPoke: https://www.svix.com/poke/Standard Webhooks: https://www.standardwebhooks.comStandard Webhooks on GitHub: https://github.com/standard-webhooks/standard-webhooksEteSync: https://www.etesync.comssot-check: https://github.com/conorbronsdon/ssot-checkMost of the Web Will Never Get APIs for AI Agents | Dhruv Batra: https://chainofthought.show/podcast/63-most-of-the-web-will-never-get-apis-for-ai-agents-dhruv-batra/Connect with Tom Hacohen: LinkedIn: https://www.linkedin.com/in/tomhacohen/X: https://x.com/TomHacohenSvix: https://link.svix.com/cotConnect with Chain of Thought host Conor Bronsdon: Newsletter: https://newsletter.chainofthought.show/Twitter/X: https://x.com/ConorBronsdonLinkedIn: https://www.linkedin.com/in/conorbronsdon/YouTube: https://www.youtube.com/@ConorBronsdonMore episodes: https://chainofthought.show Thanks to Svix, presenting sponsor of season four of Chain of Thought. Svix delivers billions of reliable webhooks for startups and the Fortune 500. Get started at https://link.svix.com/cot. Qualified startups will get $12,000 in credits, $50,000 for YC companies.

    What Breaks When Agents Become the Customer | Svix CEO Tom Hacohen
5
out of 5
43 Ratings

About

AI is reshaping infrastructure, strategy, and entire industries. Chain of Thought is the podcast where builders reason through what's changing. Host Conor Bronsdon sits down with the engineers and founders shipping AI in production to get past the hype into what's working and what isn't. Episodes cover model infrastructure, inference, agent frameworks, evaluation, and developer tools. Guests have come from NVIDIA, Google DeepMind, AMD, Databricks, Vercel, and more. Every episode carries a full transcript and show notes at chainofthought.show. New episodes weekly. Conor Bronsdon is an independent consultant and angel investor in AI infrastructure and developer tools. He led technical ecosystem at Modular, acquired by Qualcomm in 2026; led developer awareness at Galileo, acquired by Cisco; and ran developer marketing at LinearB, where he was GM of the Dev Interrupted podcast and community. Views expressed by the host and guests are their own.

You Might Also Like