goodthings

[goodthings @Goodthings]

https://t.co/nN96wxLkFl

[goodthings @Goodthings]

https://t.co/ot0Km4SCDt

[goodthings @Goodthings]

tldr: every agent needs an outcomes manager.

[Jaya Gupta @JayaGup10]

https://t.co/eUp5waUIwl

[goodthings @Goodthings]

@azaztrader01 As ideas get cheaper, judgement gets more expensive.

[goodthings @Goodthings]

AI commodifies exploration.

[goodthings @Goodthings]

@sandeepnailwal Based

[goodthings @Goodthings]

@smashleybelle Significantly. Likewise, Ash 🤙

[goodthings @Goodthings]

When exploration becomes free, judgment becomes scarce.

[goodthings @Goodthings]

@Tyler_Did_It It always goes wrong.

https://t.co/wIWDkcEgE9

[goodthings @Goodthings]

https://t.co/nN96wxLkFl

[goodthings @Goodthings]

@ujjwalscript I would submit that it's not technical debt, but cognitive debt that is going to lead to the nightmare.

We saw a glimpse of that this week with the internal Amazon meeting around the outages that occurred as a result of AI-generated code.

[goodthings @Goodthings]

Context Loops is free and open source.

Docs and framework: https://t.co/agBzZHElMx

If you're building with AI agents, this shows how to structure project context so work carries across sessions instead of resetting every time.

[goodthings @Goodthings]

A lot of people using AI right now aren't becoming more productive.

They're just spending tokens.

They open a new chat, re-explain their project, get some output, and start fresh the next time. It feels like progress. It isn't.

I know because I did this for months.

Most people miss two fundamental constraints of LLMs: they don’t remember anything between sessions, and even within a session their memory is limited by the context window.

Every new chat starts from zero. No memory of your decisions. No awareness of what you already tried. No understanding of what your goals are.

The result is chaos.

The AI contradicts itself.

It re-suggests ideas you rejected last week.

It feels like onboarding a new hire every morning who never read the docs.

After eight months building a production app with AI and getting this wrong more times than I'd like to admit, I stopped blaming the tools and started designing a solution.

The result is Context Loops, an open-source framework that structures your project context so AI agents can actually work across sessions.

Tickets. Session logs. Sprint records. Decision logs. All living in your repo, all feeding into every future session.

Your work carries over instead of disappearing.

The shift it creates is simple: you stop being a passive user of AI and start orchestrating it.

If you're building with ChatGPT, Cursor, Claude, or Copilot and your sessions feel like they reset every day, this is the fix I wish I had eight months ago.

The framework is free and open source. Full docs in the comments if you want to try it.

[goodthings @Goodthings]

The Amazon internal engineering review that leaked this week is the first of many signals that the AI narrative is way ahead of reality.

Context is all you need (but context ain't cheap).

[goodthings @Goodthings]

This is such a great analogy and sage advice from @KiteVC

Machines can't replaces humans...yet. And they may never fully be able to. However, they will replace many of the tasks we have to do today.

Learning fundamentals allows us to be creative. And creativity is something that AI is going to have a very difficult time beating humans at. Never stop learning ⚡️

[goodthings @Goodthings]

If you've ever had to train someone to takeover your job as you leave for a new one you will understand more about the challenges of AI taking over entire industries than any think piece ever could.

[goodthings @Goodthings]

Full breakdown of why this is a fundamental constraint of modern AI and is likely to keep happening — and critically, what teams can do about it:

https://t.co/urU5VGRVqI

[goodthings @Goodthings]

@Conquistador_86 These people gonna reverse course by Q3 earnings.

[goodthings @Goodthings]

“Don’t worry about token costs” is the AI bro’s version of “never stop clicking”

How’d that work out for the meme coin bros?

[Reid Hoffman @reidhoffman]

.@NotionHQ's Founder, @ivanhzhao: Stop trying to save costs when you're implementing AI. It's the easiest way to lose the race.

More on how Ivan thinks about staying ahead of the AI curve:

[goodthings @Goodthings]

@westoque @reidhoffman @NotionHQ @ivanhzhao This.

[goodthings @Goodthings]

@BryanXBT Send me a link to his work 🙏

[goodthings @Goodthings]

A lot of people using AI right now aren't becoming more productive.

They're just spending tokens.

They open a new chat, re-explain their project, get some output, and start fresh the next time. It feels like progress. It isn't.

I know because I did this for months.

Most people miss two fundamental constraints of LLMs: they don’t remember anything between sessions, and even within a session their memory is limited by the context window.

Every new chat starts from zero. No memory of your decisions. No awareness of what you already tried. No understanding of what your goals are.

The result is chaos.

The AI contradicts itself.

It re-suggests ideas you rejected last week.

It feels like onboarding a new hire every morning who never read the docs.

After eight months building a production app with AI and getting this wrong more times than I'd like to admit, I stopped blaming the tools and started designing a solution.

The result is Context Loops, an open-source framework that structures your project context so AI agents can actually work across sessions.

Tickets. Session logs. Sprint records. Decision logs. All living in your repo, all feeding into every future session.

Your work carries over instead of disappearing.

The shift it creates is simple: you stop being a passive user of AI and start orchestrating it.

If you're building with ChatGPT, Cursor, Claude, or Copilot and your sessions feel like they reset every day, this is the fix I wish I had eight months ago.

The framework is free and open source. Full docs in the comments if you want to try it.