⚡ AI That Gets Sh*t Done — From Zapier’s Playbook to LLM Cage Fights

The Byte’s Bits

♟️ AI Chess Smackdown: When LLMs Throw Down in the Digital Cage Fight

😱 Trello’s Redesign Goes Down in Flames

🤖 Developers, Reinvented: From Coders to AI Conductors

⚙️ How Zapier Quietly Built an AI Powerhouse

🏎️ What happens when you tell engineers to learn how to race cars?

Google’s new Kaggle Game Arena is turning AI evaluation into a spectator sport, starting with a high-stakes chess tournament featuring the world’s top LLMs. The semifinals just wrapped, and the stage is set for a nail-biting final between OpenAI’s o3 and xAI’s Grok 4.

🕹️ From Benchmarks to Battles: Instead of static tests, Kaggle Game Arena uses live, competitive games to test reasoning, planning, and adaptability.

🤝 Big Names, Bigger Brains: Eight heavyweights entered—Google’s Gemini, OpenAI’s o-series, Anthropic’s Claude, xAI’s Grok, DeepSeek R1, and Moonshot AI’s Kimi.

🔥 Semifinal Drama: o3 swept o4-mini 4-0 with a 12-move masterpiece, while Grok 4 edged past Gemini 2.5 Pro in a tense Armageddon tiebreak.

🎥 Streamed Like Esports: Commentary came from Hikaru Nakamura, Levy “GothamChess” Rozman, and even Magnus Carlsen.

Kaggle Game Arena isn’t just testing AI it’s making it entertaining. And with Grok 4 and o3 clashing in today’s final, this could be the most-watched AI chess game ever.

This is so much better than looking at boring chart comparisons on which model is N% at x or y. Now we have the MMA version of LLM testing.

Design is hard, and direct user feedback can be brutal. However, Trello’s new design may have thrown gas on the angry mob fire. I used to be a big fan of Trello until Atlassian took over.

Trello’s latest overhaul was meant to refresh the platform, but for many longtime users, it’s a total disaster. Complaints about clunky layouts, strange fonts, and missing features are flooding forums, with some even calling it “the Windows 8 of project management,” or the worst UX rollout in history, comparing it to the Sonos app update. Ouch!

🆘 Feature Fallout: Popular functions like quick comment-adding are gone, making workflows feel slower and less intuitive.

👥 Users in Revolt: Feedback ranges from “truly terrible” to “horrible new UI,” with veteran Trello fans mourning the old design.

🎯 Strategic Shift: Atlassian says the redesign is part of a push to reposition Trello as a personal productivity tool, steering heavy team users toward Jira.

💸 The Price of Change: For teams, the shift could mean higher costs if they’re forced to migrate to Jira’s pricier tiers.

Trello’s risky revamp shows how changing a beloved product can backfire fast, especially when loyal users feel left behind.

The CEO of GitHub, Thomas Dohmke, wrote a courageous blog post about the future of developers when it comes to AI. He answers so many essential questions? Will developers exist in the future? Yes, but the role will be completely redefined.

The role of the software developer is evolving fast. Instead of just writing code, top devs are now directing AI like a creative partner, saving time, boosting output, and unlocking new ways to build.

📍 Stage 1 – The Skeptic: AI feels gimmicky, but worth a quick test for code completion.

🧭 Stage 2 – The Explorer: Using AI to debug, generate snippets, and speed up small tasks.

🤝 Stage 3 – The Collaborator: Working alongside AI inside IDEs to co-create entire features.

🎯 Stage 4 – The Strategist: Orchestrating multiple AI agents, delegating work, and verifying output—becoming the “Creative Director of Code.”

This shift doesn’t spell the end for developers it’s a level-up. With AI in their toolkit, coders are moving from doing the work to designing how the work gets done.

The hardest thing about AI is adoption within a company. Sure, some people will adapt to it right away, others will scrape the surface, but a large majority never even log in to try it. Change is always hard, but even more so within an established organization.

Zapier didn’t just bolt AI onto its automation platform it rewired how the whole company works. From internal workflows to customer-facing features, AI is now baked into the core of Zapier’s product and operations.

🧠 AI Everywhere: Zapier infused AI into internal tools to speed up support, marketing, and engineering — cutting manual work across the board.

🤖 AI-Powered Features: Launched “AI Actions” letting users connect ChatGPT with 6,000+ apps to create workflows that run themselves.

📈 Faster Than Ever: Internal teams now use AI for code generation, copywriting, and debugging, shrinking timelines from weeks to hours.

🛠️ Experimentation Culture: Zapier’s approach? Ship fast, learn faster, and keep what works.

Zapier’s AI rollout shows how AI isn’t just a feature, it’s a multiplier. The companies that figure this out will move at a speed their competitors can’t match.

The Zapier playbook on AI adoption is really an inspiring read.

To follow on to this, I just heard the podcast from My First Million this week where they interviewed the founder of Zapier, Wade Foster. Wade walks through how he goes from concept to implementation using Zapier. Great video, with hands-on use cases.

If you are into cars, this is a great video. Chevrolet Corvettes set out with a goal to be the fastest American production car to lap around the infamous Nürburgring in Germany. Not just any circuit, but one of the longest race tracks in the world, at 12.9 miles long (20.8 kilometers).

Here’s the twist: Chervrloet didn’t just go out and hire the best drivers; they used the engineers who built and designed the car. It's amazing to see the extensive testing, engineering, planning, and simulation time the drivers went through before ever hitting the track.

Makes me want to get back into racing again…