Kent Beck 🌻

I am continuing work on "Tidy First? An Exercise In Empirical Software Design" as a @substack . No paid tier yet, but that's where book chapters will be appearing. Sign up here: kentbeck.substack.com.

All these “30% unemployment for new grads” predictions make a critical false assumption—they assume the number & composition of employers remains static. We’re going to see an explosion of new ventures, some of which will turn into new employers. My gut is that these new employers will more than absorb the people laid off by current employers.

@jrzabott tidyfirst.substack.com/p/idea-to-impa…

Hey @KentBeck , I think of you a lot as a voice and a mind I respect and look up to. I was talking to Claude and(screenshot). I've felt a little shitty and happy at the same time. I was wondering if you ever felt like this, knew somebody like this or wrote about handling it?

A new podcast, hosted by yours truly, coming soon...

@PeterJThomson @staysaasy @unclebobmartin Build it, even a little bit.

@staysaasy Hey @unclebobmartin & @KentBeck Can we bring back XP, or multi-player somehow into modern work? It might need new tooling, but there has to be a human-to-human version of modern Pair Programming? Maybe like a group chat mode for Agentic Programming?

Has anyone tried pair prompting? I feel like that should be a thing. Pair coding was a common practice where one person would code and another would watch, and they’d share ideas and work together on a problem. I think you could do the same thing with AI coding to great effect. But I’m not sure I’ve heard anyone do it. Imagine rapidly iterating on QA, checking results, switching between agents rapidly as a team.

@johncrickett @bygregorr Interesting question--all the tests or one test at a time? Nobody knows. Yet.

@bygregorr @KentBeck I tend towards Test-First now, writing all the tests I think of / know first.

TDD changed how I write software. Agentic engineering is changing it again. I don't think the classic red-green-refactor loop of TDD makes sense when an AI agent is writing the code. I've been a proponent of TDD since I read @KentBeck's XP Explained. I roughly followed what he describes as "Canon TDD": 1. Write a list of the test scenarios you want to cover. 2. Turn exactly one item on the list into an actual, concrete, runnable test. 3. Change the code to make the test (and all previous tests) pass, adding items to the list as you discover them. 4. Optionally refactor to improve the implementation design. 5. Until the list is empty, go back to #2. I say roughly, as I was never good at sticking to #2. I'd turn two or three test scenarios into runnable tests at a time, often using table-driven or parameterised tests. This process adds value when a human is writing code. But when an AI agent is doing the writing, I think Test First Development is currently a better approach. We can define all the test cases, let the AI write all the tests, then let the AI write the code to pass them. Doing each step one at a time would be a much more expensive way of using the agent. There would be far more tool calls, far more roundtrips to the inference engine, and it would consume a lot more tokens. Then there's the difference between the agent and an engineer. The agent isn't necessarily going to refactor like a human engineer would. I don't think we lose much from this change. We can still guide it after each feature to evolve the design. We can still iterate quickly and get fast feedback on the design and functionality. Are you still doing red-green-refactor when the agent is doing the typing?

@kurko @GergelyOrosz Nope. People are people. Genies are not.

@GergelyOrosz @KentBeck Can vibe coding be categorized as pair programming? 🙃 I think it can, right?

Amusing - I'm hearing a comeback of XP (Extreme Programming) practice. In the early 2000s these used to be popular (championed by @KentBeck). Then died down. They are now surging again. XP practices like small releases, frequent integration, and constant customer input!

@WomanCorn @rkobylinski @martinfowler I've been working on a project called Wait, What? to address exactly this question.

@rkobylinski @KentBeck @martinfowler I also need the ability to see that this is the change that was made without reading all the diff's and to confirm that the change i'm looking at is in fact just a rename

We should revisit automating safe refactorings not only in IDEs but also via CLI or MCP. Why use a risky, non-deterministic LLM to do a deterministic tool's job? @KentBeck @martinfowler

I've spent the last year deep in augmented coding — how genies change what programmers do & don't do. Now I'm curious what that looks like in govtech, where shipping isn't theoretical. Speaking at @Rise8's ShipSummit, Park City, March 31 - April 2.

@marcopesani @Linus__Torvalds @ID_AA_Carmack Who wins the debates?

If it must be slop, make it fine slop. Say hello to the /review command, that makes @Linus__Torvalds , @ID_AA_Carmack and @KentBeck to debate on EVERYTHING

What if we were using the same words to talk about software development but the meanings were opposite?

@dafydd_rees We should think about why that might be, what part of it we're responsible for. Not all, but some.

@KentBeck I went to a product design meetup and all I heard was ways of offshoring developmemt work, low code, no code… anything to avoid paying those hateful developer meatbags for an honest day’s work. They seem to hate us devs.

"AI value = headcount reduction" is a pinhole view. What about 3x the customers with the same team? Shipping in weeks instead of months? Experiments you never would have tried? Cost reduction is legible. But it's not the whole picture. Read more on my Substack post, sponsored by @resolveai open.substack.com/pub/tidyfirst/…

There's a gap between finishing one feature and starting the next. There's pressure to fill it immediately. That gap is where you work on your futures—the things that make the next feature easier.

Genie Sessions: me vs the AI, live and unfiltered. Always ask the genie. Every dumb question, just ask. You never know if the answer's gonna be fantastic or disaster. Full sessions on YouTube: buff.ly/xOJSFPD

Every time a product I use gets worse I imagine what person's metrics just got better.

@slava_imeshev We would concretely experiment more often. "I dunno. Build one & see." Our ratio of experimental "successes" to failure would drop, so we would throw away a larger percentage of those experiments.

@KentBeck @KentBeck, I'm super curious, what are your thoughts on this: imagine the state of art where going from idea to working code in production takes nearly zero time, let's say 10 seconds. What changes?

I know we're in the "horseless carriage" stage of augmented development & it's too soon to know what the "real" name is of this new activity, but... ...the genie suggested "wishcraft" & I don't hate it.

@dibley1973 @jesusmnavarrol @coderabbitai But maybe there's a better way to spread knowledge.

@jesusmnavarrol @KentBeck @coderabbitai Pull request reviews are DEFINITELY worth it. If nothing else it spreads knowledge.

Code review made sense when humans wrote code at human speed. Now I'm on a beach with a genie, generating faster than anyone could review. The old model broke. What replaces it? open.substack.com/pub/tidyfirst/… New post, sponsored by @coderabbitai

@swe_in_exile @coderabbitai You don't have to do anything today. Wishing you peace.

@KentBeck @coderabbitai I don't want to live in this brave new world. I simply don't. Nothing to drive me deep into a depression right on New Year's Eve than Hashimoto, Beck and DHH writing approximately the same in a span of days.