Jarel

Head of AI Transformation - Helping great teams do great things. 🚀 - Personal account, views are my own.

[Jarel @JarelRemick]

@veztek Great work! I'll have to try this out next week.

[Jarel @JarelRemick]

Wait... do we get these when @thsottiaux pushes the reset button? 🤠

[OpenAI @OpenAI]

We heard you wanted to use Codex rate limit resets on your own time.

Starting today, we’re rolling out the ability to save rate limit resets to use later.

We’re starting Go, Plus, Pro, and Business users with one free reset:

[Jarel @JarelRemick]

@thsottiaux It's a lot more fun this way

[Jarel @JarelRemick]

💯

[Garry Tan @garrytan]

Tokenmaxxing done right is outcomemaxxing

[Azeem Azhar @azeem]

A general-purpose technology can sit inside firms for ages before we see the results. It happened with electricity, it’s happening with AI.

[Jarel @JarelRemick]

Awesome! Been hoping this would come soon.

[Guinness Chen @guinnesschen]

If you ever get tired of managing your Codex threads, just let Codex manage itself! Codex can now create threads, search them, organize them, pin the important ones, and spin up worktrees for parallel tasks.

[Jarel @JarelRemick]

The dream

[Luis Mejia @l_mejiaC]

GPT-5.6-pro-spark

[Jarel @JarelRemick]

@_simonsmith Using https://t.co/xAgdpvqYDw for this, but it'd be nice if it was a built in feature!

[Jarel @JarelRemick]

@BaluGorade Largely nonsense. AI is a tool, if it's used poorly it's expensive. Used wisely it's more cost effective. Knowing the difference and how is currently expensive (humans).

[Jarel @JarelRemick]

If you can't show ROI, that's poor practice and discipline. Tell your staff to speed run tokens without guardrails and this is what happens.

[Karthik Hariharan @hkarthik]

The same conversation is happening across tech right now and many of us saw it coming. Tokens got burned for millions of dollars without any real significant ROI to show for it.

The largest accounts with OpenAI and Anthropic subs have FDEs being thrown at them to retain the subscriptions as the first year hype winds down. The rest will probably slowly churn.

I think AI is here to stay, but the monetization model will undergo a lot of changes over the next 18 months.

[Jarel @JarelRemick]

Great work 👏 Throwing https://t.co/gVTNOyiy6G at it.

[Ronan Berder @hunvreus]

From the "AI-Augmented Software Development Manifesto": https://t.co/VGjpp5X9Xe

[Jarel @JarelRemick]

Just use caution with the framing. Classic bias to conclusion without real cause validation and understanding. Look at ROI, harness practices, engineering fundamentals, etc.

[Brandon Carl @brandonjcarl]

Very interesting data across 2,444 companies.

82% of tokens are spent on AI-generated bugs, rework and review friction.

[Jarel @JarelRemick]

💯👏

[Ronan Berder @hunvreus]

Lots of "AI influencers" out there with god-awful takes. It's embarrassing. Look, we're all trying to figure this thing out, but it's not magic:

1. Ship real software, not AI theater. LOC count, agent swarms, and token burn are not progress unless users get something reliable.

2. Let requirements harden from evidence. Use AI to discuss, prototype, and test assumptions before locking the plan.

3. Treat code as cheap and understanding as expensive. Deliver increments quickly, throw away bad first passes, rewrite, and refactor aggressively.

4. Keep humans and AI in a tight loop. AI generates options quickly; humans supply taste, domain judgment, trade-offs, and the ability to say "no."

5. Build around people with taste and ownership. They should set clear boundaries, trust AI to execute within them, and have the standards to redo, undo, or throw away the result.

6. Prefer dialogue over prompt bloat and rigid process. Talk through the problem, agree on direction, then write down only what needs to persist.

7. Reliable, tested software is the measure of progress. Abundant code, docs, prompts, traces, and tool output are useful only when they improve the system.

8. Sustainable pace is non-negotiable. AI makes nonstop work seductive: late-night agent babysitting, context spirals, and fake urgency. A process that burns people out or degrades judgment is broken.

9. Technical excellence matters more when generation is cheap. Understand the architecture, probe the output, push back hard, and review security before plausible broken systems become production systems.

10. Simplicity is the art of refusing generated complexity. AI can generate code, dependencies, abstractions, and documentation faster than you can own them.

11. Good systems emerge from iteration and pruning. Prototype, learn, rewrite, document, and refactor. Do not expect unsupervised agents to develop taste or conviction.

12. Use AI to amplify agency, not replace it. It should help you learn faster, explore further, and spend more time where your taste and judgment have the highest return. If it makes you passive, you are using it wrong.

[Jarel @JarelRemick]

@mattshumer_ @nickbaumann_ Make it a skill and you’ll get better results. 👍

[Jarel @JarelRemick]

@karrisaarinen It's a systems design problem most businesses don't understand and aren't built for. The bigger the org, the more the lag because the change effort is higher. Small orgs are seeing results, but they don't make the big headlines.

[Codie Sanchez @Codie_Sanchez]

I look at this at least 1x a month.

[Bnaf.OG | 🟧 @bnafOg]

If the Microsoft/Claude Code reports are directionally right, the lesson is not “AI tools are too expensive.” It’s that coding-agent procurement needs cost per merged fix, rollback rate, and reviewer time—not seat count or demo velocity.

[Jarel @JarelRemick]

Parallel agents without context switch costs requires harnesses you can trust, and higher-level delegation (goal vs task). You build your environment to reduce your context switching, then you can scale.

[Tyler @rezoundous]

Running 3 parallel Codex sessions and I'm already feeling unproductive with the context switching. How is anyone running more parallel sessions?

[Jarel @JarelRemick]

@charliermarsh I'd get many more app and solution ideas being built out in parallel with /goal I probably try to avoid hitting weekly rate limits more than I should. Even just 50B a month would be amazing!

[Jarel @JarelRemick]

@jturntdev For main threads? Never (but Codex is a bit clunky with that). Custom sub-agents make it easy to farm out effort-specific work though.