Nice. Do you think customers really want this for the pipeline integration?

Seattle startup Actual AI raises $3.2M to build AI agents for engineering managers | John Kennedy Exciting day for our team! This has been a long time coming, but we are just getting started! Thanks Taylor Soper for the great article!

Big day for our team: www.linkedin.com/posts/johnak...

It is!

Are you doing any signing/reading outside of worldcon?

Wrong guy

They are the best!

Yay!!

Wrong John Kennedy

We’re here.

Yes! We are using @actual.ai , which is our commercial product. Happy to talk you through it if you are interested. john at

*last year!

@pfrazee.com played a vital role in enhancing the platform's security. They fixed a critical bug in mute word handling, ensuring more reliable content moderation. They also removed the client-side limit on labelers and tackled auth scope challenges for account migrations.

@bnewbold.net spent a lot of time fixing typos in descriptions and comments, showing a keen eye for detail. This developer also contributed to significant features, like allowing colons in record keys and enforcing stricter rules for repository commits.

@hailey.at focused on streamlining the application's logic and testing. They simplified code, implemented detached quotes and hidden replies in threads, and added quoteCount to the embed view, proving their ability to clean up code and enhance user experience.

@esb.lol was a versioning virtuoso. Seemingly obsessed with ensuring the platform stayed fresh, they introduced new features such as video embeds, quote aggregations, and improved mute word handling. They also tirelessly hunted down bugs related to avatar resolution and other issues.

@matthieu.bsky.team dedicated their time to improving the security and robustness of the OAuth and HTTP processes. They spent their time enhancing validation, managing tokens, and fixing various minor issues, proving their knack for finding the source of errors and implementing precise fixes.

@divy.zone 's coding journey this year could be summed up as "cleanup crew extraordinaire." They seemed to find great joy in tidying up the codebase. This developer also fixed crucial bugs related to blob deletion and profile views, saving users from potential headaches and application crashes.

@foysal.it was on a mission to fortify the platform's security. From adding threat signatures and refining OAuth validation to enhancing moderation tools and implementing rate limiting, they left no stone unturned in their quest to make the platform a fortress of digital safety.

@dholms.xyz tirelessly worked on everything from feed interactions and thread mutes to fixing critical bugs and improving performance. They were the master of all trades, juggling multiple projects and keeping the codebase in tip-top shape.

Wrapped.dev Your Repos' 2024

Love the work done by the @atproto.com team this year! Here's a fun little analysis with the top 8 contributors:
www.wrapped.dev/repo/bluesky... 🧵

The problem with most pitch critiques is that they don't recognize that there is one pitch for investors and one pitch for customers. Any pitch can be destroyed if you judge it on the alternate target.

Supabase.

New year, new static generation framework.

31K out of 50K. I need to spend much more time on this.

Wrapped.dev Your Repos' 2024

On the 31st, we hit third on ProductHunt with wrapped.dev! We’ve processed over 191K commits so far! Request your favorite open source repo to see a fun AI generated summary of 2024!

I now think of these words all over the place including customer request tagging, issue tagging, moat themes, customer/investor updates, etc.

This one is particularly useful in thinking about roadmap. For each of our agents, we have to balance Perception, Reasoning, Decisioning, and Tooling.

"[An AI agent is] an intelligent software system designed to perceive its environment, reason about it, make decisions, and take actions to achieve specific objectives autonomously" from Rudina Seseri at Glasswing.

What exactly is an AI agent? | TechCrunch Regardless of how they're defined, the agents are for helping complete tasks in an automated way with as little human interaction as possible.

For me, AI agent has meant autonomy (does things without asking) and agency (the things it does have an effect on the world, not just answering questions), but today I found a better definition reading this article: techcrunch.com/2024/12/15/w...

Apparently we are going to use the the term "agent", an already ridiculously overloaded term, for the defining technology of 2025.