iiinova — where your digital self keeps evolving

Tom Johnson @tomjohnson3.bsky.social · 29/07/2026

I know leaders who live in constant fear that a business-killing bug is lingering in some unread, AI-generated corner of the codebase. Grounding non-deterministic AI output in deterministic input is just one layer of the solution.

000

Tom Johnson @tomjohnson3.bsky.social · 29/07/2026

I think it’s emotionally exhausting to manage an infinitely hungry AI workforce while knowing there is no longer direct human ownership over every line of code.

010

Tom Johnson @tomjohnson3.bsky.social · 23/07/2026

Tokenmaxxing was always going to end. Goodhart’s Law doesn’t fail. What replaces it is the more interesting question.

030

Tom Johnson @tomjohnson3.bsky.social · 22/07/2026

Had a demo call earlier today where the CLI wizard set the live demo up in under a minute. The user's reaction said it all. Very proud of the team for building something that actually works that smoothly. That kind of experience doesn't happen by accident.

000

Tom Johnson @tomjohnson3.bsky.social · 21/07/2026

Hundreds of open PRs, reviewers on autopilot, the same bug fixed three different ways by three different agents: something is off upstream.

000

Tom Johnson @tomjohnson3.bsky.social · 21/07/2026

Show me a team's review queue and I'll tell you how they're using AI. Reasonable revert rate, reviewers who feel like they're adding value: coding agents being used well.

120

Tom Johnson @tomjohnson3.bsky.social · 16/07/2026

The truth is that we've accelerated code generation without evolving how we verify it or who owns it. Teams are producing more than they can reason about. Better abstractions, testing strategies, and system boundaries aren't optional improvements anymore. Without them, none of this scales.

100

Tom Johnson @tomjohnson3.bsky.social · 16/07/2026

I recently posted a screenshot of an AI-generated PR with 50k lines of changes. The comments exploded with opinions on PR best practices. Almost nobody stopped to ask whether the review process itself still makes sense. It's like walking past a burning building to debate the exit sign.

100

Tom Johnson @tomjohnson3.bsky.social · 14/07/2026

Finished doesn't mean closed. One of my favorite things about the Multiplayer web dashboard is that you can reopen any session and keep going in the same thread the agent used the first time.

011

Tom Johnson @tomjohnson3.bsky.social · 14/07/2026

I think part of the answer here is to ground the model's nondeterministic behavior in deterministic data (designed for AI tools).

010

Tom Johnson @tomjohnson3.bsky.social · 09/07/2026

We’re at a point where we need to start rethinking of how we work with these tools: what data we prepare for them, how we select which model does which job, and how much we trust them to reason rather than trying to pre-load the answer.

100

Tom Johnson @tomjohnson3.bsky.social · 09/07/2026

This is a quote from Cat Wu, Head of Product, Claude Code, 👇 What makes this interesting to me, beyond practical advice, it’s what it signals about the maturity curve of working with AI.

200

Tom Johnson @tomjohnson3.bsky.social · 07/07/2026

A bloated review queue full of AI-generated PRs isn't a sign that your team is moving fast. Developers who are addressing this problem are taking a step back and thinking about the whole system: what data agents work from, how duplicate issues get caught, what it takes to make a PR worth merging.

020

Tom Johnson @tomjohnson3.bsky.social · 03/07/2026

If you’re using coding agents, your output numbers (epics completed, task throughput, PR merge rate, …) may look good right now. But those are not the metrics your manager is watching. Every single one of those numbers comes with an asterisk: production quality is declining.

210

Tom Johnson @tomjohnson3.bsky.social · 30/06/2026

Telemetry built for AI coding agents captures full-fidelity, pre-correlated session data at the source, so an agent isn't reconstructing a failure from sampled traces or scattered tools.

120

Tom Johnson @tomjohnson3.bsky.social · 30/06/2026

Developers don't want more dashboards to stare at or more alerts to triage. They want the thing that broke, to fix itself. So we don't have to go through this flow of 👇 five tools, eleven steps and four hours just to fix a bug.

100

Tom Johnson @tomjohnson3.bsky.social · 24/06/2026

These look like 3 different problems but they share a common root: AI tools working with incorrect, sampled, or missing runtime data. The teams making progress are asking hard questions about how they collect, correlate, and deduplicate data before it ever reaches an agent.

100

Tom Johnson @tomjohnson3.bsky.social · 24/06/2026

880+ engineering leaders were asked what challenges their organizations have experienced with AI in the past 6-12 months. 49% flagged security concerns. 48% are dealing with hallucinations. 23% reported an increased need for debugging.

100

Tom Johnson @tomjohnson3.bsky.social · 24/06/2026

The “10x engineer” was always a simplification and I was very pleased when it started to fade. Then AI coding tools arrived, and it came roaring back. Software engineering is a team sport. The engineers who create the most impact are the ones who multiply the capacity of the people around them.

100

Tom Johnson @tomjohnson3.bsky.social · 20/06/2026

Yes. Session-based, unsampled runtime data collection is very cool 😎

000

Tom Johnson @tomjohnson3.bsky.social · 18/06/2026

And trust matters. Multiplayer runs next to your coding agent and captures full-stack production data. Now you can verify exactly what it does.

000

Tom Johnson @tomjohnson3.bsky.social · 18/06/2026

The root cause of PR slop is the same problem Multiplayer is built to solve: agents working with insufficient, sampled, or incorrect data. Better runtime data leads to better outputs, both in production debugging and across every AI coding workflow.

230

Tom Johnson @tomjohnson3.bsky.social · 18/06/2026

I won't pretend the current open source environment made this decision easy, between PR slop and declarations that 'open source is dead,' the headwinds are real. But we never stopped believing in this community, and launching into the headwind feels more honest than waiting for calmer weather.

100

Tom Johnson @tomjohnson3.bsky.social · 17/06/2026

1 in 3 engineering managers is thinking about going back to IC. 37% say their hands-on technical work has increased in the last 12 months. AI made building fast enough to compete with managing again. @leaddev.com latest report puts numbers on what a lot of us have been feeling.

100

Tom Johnson @tomjohnson3.bsky.social · 16/06/2026

The latest Jellyfish's AI Engineering Trends research shows token consumption climbing sharply, but PR revert rates remain roughly flat across all AI adoption tiers, hovering in the same 0.45%–0.65% band they were in a year ago. More tokens per developer hasn't bought fewer reverted PRs.👇

110

Tom Johnson @tomjohnson3.bsky.social · 12/06/2026

This meme is a lie and we all know it 👇 Human PR review was always a bottleneck. We merged things we hadn't fully read and shipped things we shouldn't have. The difference now is that coding agents made it worse with PR slop.

130

Tom Johnson @tomjohnson3.bsky.social · 11/06/2026

First time at #KCDNewYork. Great event, easy to meet people, fantastic food. Met @salaboy.com, newly elected CNCF TOC member and author of Platform Engineering on Kubernetes. Already at the top of my reading pile. See you at the next one. 🙌

Me, Tom Johnson and Mauricio Salatino at the Canal St. subway station

Platform Engineering on Kubernetes book by Mauricio Salatino

Platform Engineering on Kubernetes book with author's signature

Me, Tom Johnson at the KCD New York event

020

Tom Johnson @tomjohnson3.bsky.social · 10/06/2026

💯 you have have to be intentional about your workflows and tool choices.

000

Tom Johnson @tomjohnson3.bsky.social · 10/06/2026

It's a valid point, but consider that many enterprise teams are locked into a contract with a specific vendor, especially if they use other tech from that vendor (e.g. Microsoft), so switching models isn't straightforward. The same goes if you're using a debugging agent like Seer or Datadog bits.

110

Tom Johnson @tomjohnson3.bsky.social · 10/06/2026

The way we verify software quality needs to evolve at the same pace as the way we generate it. Keeping a broken review process and hoping better tools paper over it isn't a strategy. Full article: leaddev.com/ai/pr-review...

leaddev.com

PR reviews were already broken. AI made it worse

AI didn’t break code review – it just made a broken process impossible to ignore, and fixing it requires layers, not a single solution.

100

Tom Johnson @tomjohnson3.bsky.social · 10/06/2026

Who wants to review this PR? 👇😅 PR reviews were already a known weak point in software development before AI coding agents arrived. They just made it impossible to ignore.

100

Tom Johnson @tomjohnson3.bsky.social · 10/06/2026

Multiplayer works with Claude Code and very soon with Codex, Copilot, and Cursor. Vendor neutrality isn't a nice-to-have. It's the only way to keep up.

000

Tom Johnson @tomjohnson3.bsky.social · 10/06/2026

Every major model release makes the same point: the frontier keeps moving. Claude Fable 5 today. Something better in six months. The most underrated decision any engineering team makes right now: building AI workflows that are MODEL-AGNOSTIC.

410

Tom Johnson @tomjohnson3.bsky.social · 08/06/2026

I'm speaking at the #DeveloperWeek conference tomorrow in 📍New York. If anyone is interesting in discussing agentic coding, debugging, or just have a coffee, find me and @bkstephj1.bsky.social ☕️

010

Tom Johnson @tomjohnson3.bsky.social · 05/06/2026

Back in NY from #LDX3 London. 3 sessions in 2 days: AMA panel, a talk on why most MCP servers are collecting dust, and moderating a peer group. The questions after my talk made every hour of prep worth it. @leaddev.com sets the bar high for developer conferences!

010

Tom Johnson @tomjohnson3.bsky.social · 04/06/2026

Full article: open.substack.com/pub/beyondru...

000

Tom Johnson @tomjohnson3.bsky.social · 04/06/2026

It’s hard not to make the connection between between AI output quality and data input. The sooner the distinction between dashboard data and agent data becomes a first-class engineering concern the sooner coding agents will actually deliver on what everyone expects them to do.

100

Tom Johnson @tomjohnson3.bsky.social · 04/06/2026

“some companies are shipping as many as 50% more defects than before.” This is 👆 from a recent DX report on AI-assisted engineering which confirms what many engineering teams have been quietly noticing: adoption is up, but quality continues to be inconsistent.

000

Tom Johnson @tomjohnson3.bsky.social · 29/05/2026

I'm speaking at the LDX3 conference next week in 📍London, UK. If anyone is interesting in discussing agentic coding, debugging, or just have a coffee, find me and @vladistevanovic.bsky.social ☕️

020

Tom Johnson @tomjohnson3.bsky.social · 28/05/2026

hehe I recently wrote an article were I called it the "rebranded BDUF". To be fair, I always thought some design upfront was necessary 😅

000