Tom

Fixing complex problems with AI | Melbourne

[Tom @MostlyOptimal]

@Codie_Sanchez I love this. Totally going to use this

[tobi lutke @tobi]

I’m thinking about banning Claude code at Shopify until they change their mind and read AGENTS.md and .agents/skills etc.

Insisting on only reading CLAUDE.md sometimes leads to split brain problems when different team members use different tools. Just unnecessary.

[Tom @MostlyOptimal]

@kunchenguid Don’t forget effort it should be CLAUDE_FABLE_5_LOW.md

[Matt Pocock @mattpocockuk]

Thanks for taking my rather too on-the-edge tweet and inviting a serious debate - I much prefer this.

I'm interested by how you define the term "coding".

I'll just @JohnOusterhout's definition of Tactical vs Strategic programming.

Tactical: the on-the-ground, day-to-day aspects of coding.

Strategic: the long-term. Codebase health, architectural design. Making the right decisions.

I agree that AI has largely eaten tactical programming. But I've seen no evidence that it can think strategically.

Ousterhout talks about some employees as "tactical tornadoes" - able to churn out astonishing amounts of work with zero eye for the future.

That's how agents feel to me right now.

I'd love to see how your observations fit into that frame.

[Tom @MostlyOptimal]

The playbook is lacking the whole North Star vision.md to prevent things drifting too far away from the core product. If it can happen all the time with people, and it will too with agents https://t.co/bAnb7p1jt4

[Tom @MostlyOptimal]

After playing around with @bot, I have to say it’s pretty damn intuitive for regular users. What a fantastic product!

[Tom @MostlyOptimal]

Might need to play around with this one.
It seems a bit more intuitive for a non-technical user. There is still an adoption challenge to convince people that they need it at that price point, so still early days.

[Grok Bot @bot]

Introducing Grok Bot, now in early beta.

Bots are AI teammates that do real work for you. They sign in to your tools, use them just like you do, and come back with finished work.

[Tom @MostlyOptimal]

@ClaudeDevs Cool feature

[Tom @MostlyOptimal]

@OpenAIDevs @awsdevelopers @cursor_ai @github @code @vercel Great to see more open standards here

[Tom @MostlyOptimal]

Now that's pretty cool. All it takes is a better harness to squeeze more out of our existing models

[Prime Intellect @PrimeIntellect]

Prime Agent is a general-purpose coding harness

On ARC-AGI-3, it scores 95.5%, surpassing the human-expert baseline, but the gain is not benchmark-specific.

We see major improvements across models when compared to their proprietary harnesses:

[Tom @MostlyOptimal]

Enjoying FSD 14.3.5. Feels more confident and natural on the roads. Speed sign recognition still hit and miss though