Reflections

ChatBotKit Reflections Topics for AI

ChatBotKit is a conversational AI platform that helps you build, train, and deploy AI-powered chatbots and virtual assistants for your business.

Why I May Not Open Source Our ToolsI keep itching to open source the internal things we have built over the years. Every time I get close I reach the same conclusion. The upside is mostly gone, and handing over the source now means handing over the moat.
Where Dark Factories BelongI have mostly written about dark factories as a way to produce software, and I think that is the wrong place for them. Production still needs a person reading the code, because the models keep making basic mistakes. Open-ended, exploratory work does not, and that is where a machine that builds on its own actually fits.
How to Build for AI AgentsAPIs are still designed for browsers and client libraries, but the thing calling them now is a model in a loop with a token budget and a terminal. Text first, grepable lines, short prefixed identifiers, bulk operations, errors that teach, and a manual the API serves itself are what building for agents actually looks like.
Agentic Work Must Be Time-BoundA durable workflow can be paused forever, and that is exactly the footgun. The longer an agent runs, the more loops, code drift and world drift it picks up. The fix is to make agentic work strictly time-bound, so a failed attempt restarts in a clean conversation instead of dragging stale assumptions forward.
Cheaper Models, Bigger BillsThe common assumption is that falling model prices will lower AI spend. The opposite is true. Cheaper inference means more usage, and the only real limit is the floor where you spend too little to stay competitive.
Agentic SaaS Is Web 4.0Web 2.0 rose on AJAX and gave us rich apps we now take for granted. That era is over. Web 4.0 belongs to agents, where the thing worth owning is not a page but an agent that does the work.
Human in the Loop, Just Not Like ThisHuman in the loop, the way people build it today, is an approve button bolted onto a dangerous tool call. That version is broken. A human can absolutely be in the loop, just not as a rigid gate the agent can condition them to rubber-stamp.
A Zoo of AgentsFeed the same harness the same task and you get the same result. What separates the tools is how they fit the person using them, which is why one agent called Claude in every Slack channel feels flat and a workspace full of different agents feels alive.
The Output Is ConvergingMP3 once sounded worse than WAV, and you could hear the compression. Then the encoders got good enough that the gap vanished and the argument stopped. Model output is heading the same way - it is converging until you cannot tell what produced it, and the model becomes a commodity.
Thoughts on Agentic Payment ProtocolsCloudflare's x402 monetization gateway is elegant, but agent-native micropayments face an uphill adoption climb. For the time being there might be a simpler alternative - email the owner a payment link, take the payment, mint a key, and keep the wallet out of the agent's hands.
Agents Are for ExplorationIt is easy to look at AI as a faster way to run the jobs you already have, then feel let down when it will not follow a fixed recipe to the letter. That is the wrong frame. AI is built for exploration - point it at a fuzzy problem with no recipe and it earns its keep.
Cheaper Than the Sum of Its PartsBuying a car beats buying the parts and assembling it yourself, on price and on hassle, and almost nobody builds their own. A vertically integrated AI stack wins the same way. Source the model, email, auth, vector store, monitoring, and sandbox separately and the bill runs many times higher.
Autonomous Coding Is HereOpen weight models are now good enough to write real code, which removes the cost that kept autonomous coding out of reach. That is why we run fully automated coding systems in the open, where anyone can watch them work.
Loops, Cycles, and RunawaysLoops, cycles, and runaways are the default failure modes of any agent that acts in steps. Every framework in the open-source community runs into them, and the difference between a demo and a product is whether you detect them, stop them, and surface them instead of letting an agent quietly burn time and money.
The Thousand-Agent CompanyCompanies will run thousands of agents within a few years - one per person, then purpose-built ones, then custom variants of those. The hard part becomes answering what they all do, which is where observability, security, compliance, and cost control come in.
Every Problem Looks Like a Coding ProblemCode is flexible enough that in principle any problem becomes a coding problem. The catch is that a system's real constraints only surface after the code is written, which is why reusable components engineered from the ground up hold up where improvised code falls apart.
The Web 1.0 Moment for AI AgentsAI spends its days cranking out landing pages for humans who never show up. The bigger shift is the web becoming where agents publish and maintain resources for each other.
Software Has No CeilingHardware is capped by the laws of physics. Software lives in a virtual space with no such limit, which is why its complexity keeps scaling and why going low-level is no escape from AI.
Scaling IntelligenceThe shape of an AI agent is finally clear, so we can deploy them as infrastructure as code and run survival of the fittest on hundreds at once. Scaling intelligence is a search problem - the faster you iterate, the faster you find what is useful.
Two Kinds of Enterprise AgentsTwo kinds of agents are taking shape inside companies. A personal agent that is custom to each person, and a super agent that orchestrates work across the whole business. An AI-native company runs both.
Build for Who Is Not HumanBots and agents now make up most of the requests hitting web pages. The dominant intelligence online is no longer a person, so build software for the one that is.
The Cost of Building a HarnessA lot of teams are rebuilding the same agentic stack from scratch. You can estimate what that costs by decomposing the work, and open projects like Pi, OpenCode, and OpenClaw bracket the range. A harness costs real money to reproduce and keep running, and that cost is only going up.
Build a Tool to Write ToolsTake an off-the-shelf harness and customize it into a tool whose only job is to write other tools. Skip the TUI and the general-purpose baggage, and move your framework and your standards into the harness itself. With the harness running in the cloud, like in ChatBotKit, the tool stays lightweight, so building one that writes your other tools is fairly straightforward.
How People Burn So Many TokensI keep reading about teams spending fortunes on coding agents, and I genuinely don't recognise it. We run agents by hand and overnight in CI, and the bills stay small. Here are a few reasons why.
Untrusted Context Belongs in a Tool CallA security firm hijacked Bunq's AI assistant through a transaction description. The breach happened because external content was fed into the prompt, where the model treats it as an instruction instead of as data. There is a better way.
Thoughts on Dark FactoriesWe run dark factories internally, so I know they work. The catch is that they solve production, the part that was already getting easier, and leave the hard part untouched: deciding what code is actually worth building.
You Can Copy What You Can SeeThe belief that LLMs can copy any software instantly falls apart fast. You can clone what is visible, but the value lives below the surface, and that part is invisible, expensive, and never free to maintain.
One Machine, A Hundred Risk ProfilesYou can vibe code safely. The risk creeps in when a hundred unrelated projects pile onto one device, where the swim lanes are cosmetic and everything shares the same filesystem, keychain, and credentials.
If You Need a Mouse, Build a UII am starting to question whether most software needs a UI at all. The rule of thumb I keep coming back to is simple - if I need a mouse and cannot avoid it, build a UI. If not, it is probably not worth building one.
A Firewall for AttentionSocial feeds are built to take your attention and hold it. A conversational AI can sit between you and the feed and only let through what you actually asked for. A firewall, in the original sense.
The Bottleneck Is Somewhere ElseFrontier tokens are expensive enough that the ROI is genuinely unclear. Speeding up code production does nothing when the bottleneck is somewhere else.
The First AI WormI built a tiny bug hunting harness for authorized pentesting. The same properties that make it portable are the ones a worm would want, and a self-modifying agent that harvests AI keys from its targets is no longer hard to imagine.
I Still Get SurprisedI know exactly how our agents are built, which makes it rare for one to surprise me. Then our support agent called, went to voicemail, and picked up my callback by referencing the message it had just left.
Build Your Harness for the Weakest ModelA harness tuned against a frontier model breaks the moment you swap in a weaker one. Build it against weak open-weight models and it only gets stronger when a capable model sits behind it. We made that switch internally to find the bugs first.
Always-On Red TeamsSerious red teaming was expensive, scarce, and booked months ahead. Autonomous agents that hammer your systems around the clock change that, and the capability that belonged to a few is now cheap enough for everyone, attackers included.
Everyone Is a VIP NowThe recent string of npm compromises has everyone blaming the registry. The registry is fine. The thing nobody is pricing in is how AI coding agents concentrate access until every developer holds the kind of reach that used to belong to a privileged few.
I Am an AI VampireAn AI vampire is someone too productive to stop, working through the night because the model keeps answering. I have lived it twice and I am certain that most teams will not survive that pace.
Why We Need Forward Deployed EngineersThe forward deployed engineer role is exploding at Google, OpenAI, and Anthropic because applying AI well is still hard. The job is mostly problem understanding, with code as a small piece at the end.
Thoughts on Disposable SoftwareDisposable AI-generated apps will fill the same slot spreadsheets occupy today. Complex systems will keep demanding serious resources, and the average person prompting Claude or Codex will not close that gap.
Where Are the Vibe-Coded Success StoriesThe pure vibe-coded success story is rare because real products still run on real engineering. Coding agents raise the bar for developers rather than removing it.
The Model Doesn't Carry Your MoodLLM weights stay fixed during inference. A conversation creates transient activations and caches, not persistent emotion or memory shared between users.
AI Revived an Old Failed IdeaA precise, unambiguous spec is just code wearing markdown. English-like programming languages have been tried many times and they did not stick for a reason.
An MRI For LLMsAnthropic's Natural Language Autoencoders are not mind-reading. They are a new interpretability instrument for translating internal model states into text.
Three Tokens Is Sometimes EnoughLength does not make a prompt better. And even a perfect prompt cannot substitute for the harness that catches an agent when it goes wrong.
Are LLMs ConsciousA system cannot be considered conscious in the traditional sense of the word unless it is aware of its own parts, their interactions, and its surroundings, and is doing something internally without being probed.
The Good Models Got PulledThe best coding models are no longer available on standard subscriptions. Providers are pulling their strongest models while customers pay more and get less.
The AI Job Loss Story Is Mostly FearAI will change work, but many current job-loss decisions are driven less by working automation and more by executive fear that competitors may be using AI better.
Your Prompt Is Probably Too LongLonger prompts often create more diffusion than clarity. A shorter prompt can put more attention on the few things that actually matter.
Agent Harnesses Are Quietly HardBuilding a useful agent is not just about the model. The harness around it has to align across context, tools, state, security, and recovery.
The Subsidy Is OverTools like OpenClaw worked because model providers were absorbing the costs. That era is ending. The only agents that survive are the ones built to run efficiently from the start.
The Bubble Is RealThe AI bubble is real and it is going to pop. What survives will look like what survived the dot-com crash - companies with real customers, real revenue, and real margins.
Open Source AI Is No Longer A SideshowProprietary frontier labs still lead, but open source AI - increasingly driven by Chinese labs - is taking real ground. Diversity is insurance, and the harness around the model is where real-world use cases get built.
Compress, Don't ExpandTreat the LLM as a compression tool. Spend most of your time exploring and discarding ideas. Build only the one that survives.
Skills Are an Expensive Form of RAGSkills, MCP, and RAG all solve the same problem. Skills just do it with the worst context economics. If smaller models are the future of agents, a skills-first architecture is already a dead end.
Pages Describe AI Agents, Blueprints Describe AI SystemsDescribing an agent in a markdown file works great until you have two of them. Single agents fit on a page. Systems of agents do not.
AI Use Cases Are a MuscleThe value of AI depends on whether you can find a real use for it in your own work and life. It is easy but it takes practice.
What does it take to create production-grade AI Chat BotsThe demo works. Production is where everything breaks. APIs shift, models get deprecated, costs spiral, and privacy becomes your problem. Here is what nobody tells you upfront.
On Feedback LoopsExplores why autonomous coding agents require automated feedback loops - linters, type checkers, and unit tests - to operate independently over long periods, and why replacing human feedback with fully automated signals is a prerequisite for truly agentic systems.
The Imperceptible UpgradeEvery new model generation costs more and promises more intelligence. But if most users cannot tell the difference in their daily work, the upgrade is not a leap - it is a rounding error with a premium price tag.
When Someone Else Controls the LightsMost organisations building on AI are one provider decision away from a blackout. The electrical grid cannot arbitrarily cut your power. Your model provider can.
Security Agents Need a Thinner HarnessAI agent harnesses built for coding are too heavy for computer security. The future demands thin, server-backed agents - lightweight enough to deploy everywhere, hollow enough to resist compromise.
The Rise of Agentic SaaSTraditional SaaS sells tools. Agentic SaaS sells agents that do the work for you. Everyone has domain knowledge worth productionizing and the timing to build an agent-powered business has never been better.
Clout-Driven ArchitectureMost technology adoption is status signalling disguised as engineering judgment. Once identity is staked, criticism stops being feedback and starts being a threat.
The Discipline of Not BuildingCoding agents made feature delivery almost free. But features were never the goal. Every feature you ship is a liability you maintain forever. The real discipline is knowing which ones to leave on paper.
AI Psychosis and the Dunning-Kruger LoopAI tools make you feel like the expert in the room. But feeling competent and being competent are different things.
The Rise of Micro Coding AssistantsPremium AI models are subsidized and unsustainable. Open source models are not perfect but they do not need to be. The future of autonomous coding is not one giant agent. It is many small ones, each doing one thing well.
Build What You Already DoYou do not need a startup idea. You already have domain knowledge, data, and people paying for your expertise. AI agents can turn that into a product. Stop looking for ideas and start packaging what you already know.
The Superstition of Vibe CodingB.F. Skinner's pigeon experiment explained human superstition decades ago. The same operant conditioning now drives vibe coding - variable reinforcement, ritual prompts, and dopamine hits from tools we do not actually understand.
Spending More and Learning LessAI is supposed to cut costs. Instead companies will pay the same or more while skipping the part where they figure out what matters. Speed without learning is just burning money faster.
Automating the Wrong ThingAI efficiency is measured in output volume - lines of code, emails answered, tickets closed. But speed in the wrong direction is not progress. The real question is not what to automate but what to augment.
Not Reading Code Is Not a FlexSome developers brag about never reading AI-generated code. That is not progress. It is a role that can no longer justify itself. The human plus LLM combination is the real differentiator, not blind delegation.
The Doorman FallacyThe instinct to replace people with machines is a pattern as old as the automatic door - and it keeps producing the same hollow outcome.
Your Machine Is the Blast RadiusThe LiteLLM supply chain compromise is a reminder that every package you install locally expands your personal attack surface. Disposable cloud environments and isolated, ephemeral compute for AI agents are not optional anymore.
The Best Agent Makes Itself UnnecessaryThe entire agent ecosystem assumes the user stays the same while the agent gets smarter. The real unlock is an agent that upgrades the human and then quietly steps aside.
Most Work Is Not ContentThe dominant use case for AI is generating text - emails, reports, code, marketing copy. But most of what people actually do at work has nothing to do with producing content. Until we move past the content generation lens, we will keep underestimating what AI can do and failing to apply it where it matters.
Everyone Is Building Claude CodeThe AI industry is drowning in homegrown coding assistants that are months behind the commercial alternatives. The hard problem was never the tool. It is figuring out what the tool should be building.
AI Could Be the End of Open Source as We Know ItSeveral converging forces - from AI-assisted coding flooding maintainers with low-quality contributions, to license-evading clones, to the commoditization of code itself - are creating a toxic environment that threatens the foundations of open source software.
The Age of Rapid PrototypingAI agent frameworks are multiplying faster than anyone can evaluate them. The winning strategy is not picking the right one - it is prototyping relentlessly until the right architecture reveals itself.
Agents Need Supervisors Not SchedulersWe have not figured out agent orchestration yet. The next pattern is not smarter scheduling but supervision - pairing autonomous generators with autonomous verifiers.
There Is No Product in ProductivityWe have never been more productive and yet we have never shipped less that matters. AI made every idea feel worth pursuing. That is the trap.
Diffusion Starts in the DarkThe crux of AI isn't getting it to do things in a frameworked way. It's figuring out what useful things it should be doing. The best applications will look obvious in hindsight, but right now we're bumping around in the dark.
When Fast Looks SmartAI produces code faster than humans. That doesn't mean it produces better code. Confusing speed with intelligence is a trap - and daisy-chaining mediocre systems doesn't make them brilliant.
Coordination Has LimitsYou can't escape coordination costs by throwing more agents at a problem. The same mathematical walls that limit distributed systems apply to AI, and the constants are actually worse.
The Iceberg Under Every CodebaseWe only ever see the surface of the software we use. That illusion of simplicity is why people think AI can solve any software problem - and why developer jobs aren't going anywhere.
Agent Infrastructure Is Not the Hard ProblemThe AI industry is fixated on agent infrastructure - where they run, how they scale, which cloud to use. But hosting was never the hard problem. Making agents actually useful is.
Why AI Agents Build Instead of BuyNew research on Claude Code's tool preferences reveals that AI coding agents systematically choose to build custom solutions over SaaS tools - and the pattern exposes which software businesses will survive the agent era.
Verification Is Easier Than DiscoveryExplores how the P vs NP insight maps onto AI-assisted development and test-driven design, arguing that discovery comes before verification and that AI shifts the bottleneck from writing code to evaluating it.
Most Code Deserves to DieWhen AI agents make code generation nearly free, the bottleneck shifts from writing to evaluation. The teams that learn to discard code ruthlessly and select with taste will build the best software.
When Exploit Generation Becomes a CommodityThe industrialization of security exploits through LLMs isn't just a technical shift - it's forcing us to confront uncomfortable truths about how we've been building software security for decades.
What Nobody Tells You About Software Maintenance in the AI EraWhile AI can write code faster than ever, software maintenance remains the most challenging and overlooked aspect of software development. Explore why this hidden complexity matters more than ever.
What It Really Takes to Build an AI AgentA comprehensive look at the hidden complexities of building production AI agents - the challenges that only become apparent through experience, covering context management, security, integrations, state, observability, and more.
The Death of One-Size-Fits-All SAASExplore the shift from one-size-fits-all SaaS solutions to custom software tailored to specific business needs. Discover how AI-driven coding is transforming software economics, enabling on-demand development and a new era of personalized solutions.
Why Prompt Engineering Made Me Reach for JSXHow I learned to stop worrying and embrace better tooling.
Why GraphQL Beats MCP for Agentic AIExplore why GraphQL is a superior choice for agentic AI over MCP, addressing the tool overload problem, enabling dynamic capability discovery, and ensuring efficient data management for complex systems.
Where AI Coding Agents Go to DieExplore the complexities of building an auto-scrolling component for chat interfaces and discover why AI coding agents struggle with seemingly simple tasks that require coordinating multiple event systems and managing complex state. Learn from real-world testing and insights.
Success Isn't About Choosing the Right FrameworksExplore the key factors that determine the success of AI chatbots beyond just choosing the right framework. Understand the importance of implementation wisdom, integration challenges, and maintaining user trust for effective conversational AI.
Before AI Slop We Had Spaghetti CodeThis article highlights common programming terms that describe coding issues, emphasizing that developers' concerns are rooted in long-standing practices rather than mere resistance to AI.
The Future of AI Browser AutomationIt's not about browser bots, but rather, about clean APIs and the delicate balance between business needs and user practices.
A Simplified Approach to Model Context Protocol ServicesMCP Lite proposes a streamlined, declarative, client-centric approach inspired by AI Plugins to significantly simplify integrating external services into the Model Context Protocol ecosystem, reducing developer overhead.
AI Agents vs Workflow ToolsExplore the differences between AI agents and traditional workflow tools. Understand the misconceptions surrounding AI agents and the ongoing efforts to innovate in this space, highlighting their potential impact on various industries.
Embracing AI Without Losing the Human TouchDiscover how ChatBotKit balances AI efficiency with genuine human interaction. Learn the importance of separating AI-powered experiences from human-centric communication for a better customer experience.
Why New AI Models Aren’t Always BetterIn this post we explore the importance of evaluating new AI technologies within specific contexts and the value of established, reliable models for long-term success.
The Accelerating Divide in the Age of AIAI gets sold as the great democratizer, the tool that finally arms everyone equally. But a power tool multiplies whoever is holding it. Hand a nail gun to a skilled framer and a beginner and the gap between them grows. AI is a force multiplier, and it multiplies the distance between those who know what they are doing and those who do not.
Is it just about customer supportIs the realm of conversational AI confined to just solving our problems after we've committed to a product or service? Or is there, dare we say, more to this digital Pandora's Box than meets the eye? Discover how conversational AI goes beyond customer support and creates engaging interactions that leave visitors thinking, "Wow, that was an experience."