How are devs at AI startups and in Big Tech using AI tools, and what do they think of them? A broad overview of the state of play in tooling, with Anthropic, Google, Amazon, and others.

… LLMs are a new tool for building software that us engineers should become hands-on with. There seems to have been a breakthrough with AI agents like Claude Code in the last few months. Agents that can now “use” the command line to get feedback about suggested changes: and thanks to this addition, they have become much more capable than their predecessors.

As Kent Beck put it in our conversation:

“The whole landscape of what’s ‘cheap’ and what’s ‘expensive’ has shifted.

… It’s time to experiment! — Read More

When I use AI in my software engineering job, I use it “on tap”: when I have a problem that I’d like to run past the LLM, I go and do that, and then I return to my normal work.

Imagine if we used other software engineering tools like this – for instance when I have a problem that I’d like to solve with unit tests, I go and run the tests, before returning to my normal work. Or suppose when I want to type-check my codebase, I open a terminal and run npm run tsc. Would that be a sensible way of using tests and types?

Of course not. Tests and types, and many other programming tools, are used continuously: instead of a developer deciding to use them, they’re constantly run and checked via automation. Tests run in CI or as a pre-push commit hook. Types are checked on every compile, or even more often via IDE highlighting. A developer can choose to run these tools manually if they want, but they’ll also get value from them over time even if they never consciously trigger them. Having automatic tests and types raises the level of ambient intelligence in the software development lifecycle. — Read More

“No vibe coding while I’m on call!” declared Jessie Young, Principal Engineer at GitLab, encapsulating the fierce debate dividing the software development world. On one side stand cautious veterans like Brendan Humphreys, CTO of Canva, who insists, “No, you won’t be vibe coding your way to production.” On the other hand, technology giants like Google co-founder Sergey Brin actively encourage engineers to embrace AI-generated code, reporting “10 to 100x speedups” in productivity.

“Vibe coding”—a term coined by AI pioneer Dr. Andrej Karpathy, key architect behind ChatGPT at OpenAI—has rapidly evolved from casual meme to industry-transforming methodology. In their forthcoming book Vibe Coding: Building Production-Grade Software with GenAI, Chat, Agents, and Beyond, technology veterans Gene Kim and Steve Yegge wade into this contentious territory with a bold claim: this isn’t just another development fad but a fundamental paradigm shift that will render traditional manual coding obsolete. — Read More

I was planning to write a nice overview on using claude code for both myself and my teammates. However, the more I experimented with it, the more intrigued I became. So, this is not an introductory article about claude code – Anthropic already released an excellent version of that. Instead:

We will be doing Serious Science™

What does that mean, exactly? Well, some of this is valuable, but other parts are a bit more…experimental, let’s say.

“Sometimes science is more art than science, Morty. A lot of people don’t get that.” – Rick Sanchez

Additionally, I wouldn’t say this is the most budget friendly project. I’m using Claude Max which is $250 a month. I’ll let you decide on how much money you feel comfortable lighting on fire.

Nevertheless, let’s not waste any more time… — Read More

AI agents can write code, summarize reports, even chat like humans — but when it’s time to actually do something in the real world, they stall.

Why? Because most tools still need clunky, one-off integrations.

MCP (Model Context Protocol) changes that. It gives AI agents a simple, standardized way to plug into tools, data, and services — no hacks, no hand-coding.

With MCP, AI goes from smart… to actually useful. — Read More

There’s a lot of modern techniques that have been developed since the original Attention Is All You Need paper. Let’s look at some of the most important ones that have been developed over the years and try to implement the basic ideas as succinctly as possible. We’ll use the Pytorch framework for most of the examples. Note that most of these examples are highly simplified sketches of the core ideas, if you want the full implementation please read the original paper or the production code in frameworks like PyTorch or Jax.

1. Group Query Attention
2. Multi-head Latent Attention
3. Flash Attention
4. Ring Attention
5. Pre-normalization
6. RMSNorm
7. SwiGLU
8. Rotary Positional Embedding
9. Mixture of Experts
10. Learning Rate Warmup
11. Cosine Schedule
12. AdamW Optimizer
13. Multi-token Prediction
14. Speculative Decoding

Read More

#devops

I have been developing Agentic Systems for around two years now. The same patterns keep emerging again and again, regardless of what kind of systems are being built.

I have learned them the hard way and many do so as well. The first project is not a great success, but you learn from the failures and apply the learnings in the next one. Then you iterate.

Today, I am sharing my system of how to approach development of LLM based applications from idea to production. Use it if you want to avoid painful lessons in your own projects. — Read More

“AI Engineering” is a term that I didn’t hear about two years ago, but today, AI engineers are in high demand. Companies like Meta, Google, and Amazon, offer higher base salaries for these roles than “regular” software engineers get, while AI startups and scaleups are scrambling to hire them.

However, closer inspection reveals AI engineers are often regular software engineers who have mastered the basics of large language models (LLM), such as working with them and integrating them.

So far, the best book I’ve found on this hot topic is AI Engineering by Chip Huyen, published in January by O’Reilly. Chip has worked as a researcher at Netflix, was a core developer at NVIDIA (building NeMo, NVIDIA’s GenAI framework), and cofounded Claypot AI. She has also taught machine learning (ML) at Stanford University. — Read More

OpenAlpha_Evolve is an open-source Python framework inspired by the groundbreaking research on autonomous coding agents like DeepMind’s AlphaEvolve. It’s a regeneration of the core idea: an intelligent system that iteratively writes, tests, and improves code using Large Language Models (LLMs) like Google’s Gemini, guided by the principles of evolution. — Read More

#devops

Google DeepMind today pulled the curtain back on AlphaEvolve, an artificial-intelligence agent that can invent brand-new computer algorithms — then put them straight to work inside the company’s vast computing empire.

AlphaEvolve pairs Google’s Gemini large language models with an evolutionary approach that tests, refines, and improves algorithms automatically. The system has already been deployed across Google’s data centers, chip designs, and AI training systems — boosting efficiency and solving mathematical problems that have stumped researchers for decades.

AlphaEvolve is a Gemini-powered AI coding agent that is able to make new discoveries in computing and mathematics. — Read More