I am offering six free 1:1 CTO coaching sessions in May!
If you the are most senior technology leader at a start-up or scale-up and have a challenge that you would like support with then please book one.
https://monkeysthumb.co.uk/book/free-coaching-session/
AI has compressed the product development cycle from months to days. But what happens after that first AI-generated MVP succeeds? In this post, I explore hidden costs and balancing strategies.
www.lucagrulla.com/posts/slop-ai/
Anthropic's latest paper, "On the Biology of a Large Language Model," offers a fascinating analysis of LLM "reasoning". Whether you believe LLMs are stochastic parrots or a new form of synthetic intelligence, it's worth reading.
transformer-circuits.pub/2025/attribu...
I think we will end up having prompt tracking aligned to commit SHA, stored in a separate file in the repo. In the case of micro steps, one commit == one prompt. In general, x prompts to achieve one commit.
OpenAI’s announced support for #MCP (ModelContextProtocol) is a game changer for the integration of AI into products. We now have the HTTP 1.0 of the AI revolution.
Next: a lingua franca for agent orchestration.
Meanwhile, GH/VS Code, serving the late majority, requires more control and checks. GH's greatest strength—market reach—has become its current limitation.
The Cursor AI vs GitHub Copilot race perfectly illustrates how market leadership could impede innovation.
Cursor is becoming the category king not through smashing product superiority but through rapid release cycles enabled by their risk-tolerant early adopters.
The security implications of LLM-first software are enormous.
I agree quality is not fully there. Still, it's only a matter of time. For 2025, I envision a process that begins with LLM (provide me with a baseline), followed by human input to refine v0 of the code (either directly or through prompts), rinse and repeat. This would be quicker for 70% of the work.
It feels like it is a good day to remind ourselves that there are many alternatives to Amazon (and their subsiduaries, AbeBooks, Audible and Book Depository) when it comes to getting books.
You can find many of them here
https://www.ethicalbooksearch.com/
They know that LLMs will become commoditised over time, and the diversification between commercial and OSS will become increasingly limited. Market dominance by API/models is not sustainable. The money is in workflows for the enterprise and becoming the de facto "personal assistant" for consumers.
I did the same a few years ago. Not perfect, but I agree it's a worthwhile alternative to Google.
Over the past 6-12 months, I've been using Perplexity more.
My take on 2025: What's going to happen with AI-powered software creation? Smaller teams for MVPs, AI pair programming to become the norm. And protocols and open standards, please!
www.lucagrulla.com/posts/AI-pow...
The Model Context Protocol(modelcontextprotocol.io/introduction) recently introduced by Anthropic is a step in that direction.
