jaaron
Joined 19 years ago
Karma
2,203
Submissions
450
Member since
Jan 18
Recent submissions
Comment · 5 days ago
Not who you replied to, but:
I use agent-shell and I love the ecosystem that's being created there:
https://github.com/xenodium/agent-shell
And for claude interacting back with emacs:
Some folks like having an MCP package. I agree that gives you a bit more control. But I live dangerously and run emacs in daemon mode and tell claude to just run emacsclient to execute elisp as needed. It works amazingly well. Claude has full visibility into my editor and together we can pretty much do anything.
In fact, I increasingly think of emacs as just a live lisp machine. It's a surface that I and the agents can interact against. It acts as a document store, database, control surface, etc.
Comment · 6 days ago
I like this, but the actual finding is a bit different. Trying to keep the water is wet makes this awkward, but I believe it's more like:
Everyone agrees that water is wet (P!=NP), but no one can prove it. In the meantime, a whole field of computer science was built on a stronger belief (ETH): that water, at its worst, soaks. Turns out, even at its worst, water only makes things damp. It's still wet. That said, this finding did produce a new measurement of wetness.
Comment · 3 months ago
Commenting late, but just want to thank you for this. I've been using it for a bit now, replacing vterm for my setup. I'll be sure to submit feedback via github.
Comment · 5 months ago
For similar reasons, I strongly prefer org-mode to markdown. I find that with org-mode and extensions (such as in-line elisp) I have a _significantly_ more powerful system. For example, specs can have tasks and roadmaps inline which reduces risk of drift. The biggest downside is, unfortunately, not enough folks are emacs proficient.
I hadn't considered HTML and I'm definitely going to try this.
Comment · 1 year ago
I was thinking the same thing: why is everyone reinventing emacs?
gnu and emacs already have a long history of cli and text friendly solutions that LLM dev agents can easily use and are trained on.
Or for structured data, just use a database. Dev agents can work with SQL just fine.
Comment · 1 year ago
Agreed that picking the right size of work is critical.
I didn't know about Kiro specs. I've been playing around with my own org-mode based approach with mixed success in keeping dev agent work tracked:
Comment · 1 year ago
Great write up. Love to see more of this topic on Hacker News! We've been building in this exact space since 2023, though we jumped directly into 3D with real-time multiplayer from the start.
My take: The "not a game" positioning feels like surrendering before the real design challenge begins. With our first experiment (Retail Mage [1]), we deliberately created a hybrid between traditional game structure and AI improvisation. We released it precisely because we were getting similar feedback from folks that an early implementation was too open ended and didn't "feel like a game." While I think we proved our point, we also learned a lot (including some mistakes we made).
That said, I _also_ think we can build _more than games_ with this technology. I think we’re edging into a new genre or medium: more improvisational and participatory than either traditional games or film. Structure still matters in this space, though. Without it you're really in the fanfic engine or writing assistant space. That’s fine! But it's a different goal.
Honestly, I'm very interested in this space. I was just speaking on a panel about this Monday at GamesBeat in LA and a few of us have a new podcast about this topic "Playing With Inference" [2] with folks like Nick from AI Dungeon as guests.
[1] https://www.jamandtea.studio/news/making-retail-mage-a-new-a... [2] https://playingwithinference.com/
Comment · 1 year ago
Last year, my startup launched Retail Mage as our first answer to the question of if there was any point in introducing ML/GenAI into the real-time gameplay loop. I hope this to be the first of a series of articles on what we learned.
Some other highlights, specific to the Hacker News folks:
- We have our own AI backend, running in AWS. - Costs were a major issue and we did a lot to lower it - We use a hybrid approach of mixing traditional game AI with machine learning - We have not just AI NPCs, but an AI world interaction model
Anyway, happy to answer questions!
Comment · 1 year ago
When I look back at it, Dugin was right back in November:
"So we have won. That is decisive. The world will be never ever like before. Globalists have lost their final combat. The future is finally open. I am really happy."
https://x.com/AGDugin/status/1854136340184490282
Dugin wasn’t exaggerating when he said, “We have won.”
Today’s White House display is exactly what he meant. A decade ago, the GOP would’ve never echoed Russian talking points. Now, they’re openly aligning with them.
This isn’t just isolationism—it’s the dismantling of U.S. hegemony from within.
Comment · 1 year ago
These are the FP8 distilled versions of DeepSeek we've started testing at Jam & Tea.
We use LLMs for real-time gameplay. We released Retail Mage last year on Steam as a tech demo of what we can do. We've found FP8 to be the current sweet spot for accuracy and performance for our use case. It's one of the many techniques we applied last year to bring our real-time inference costs down by 3 orders of magnitude to make Retail Mage releasable.
Anyway, we're just sharing this with anyone who finds it useful.
You can read a few more note by our ML engineer, Yudi, on LinkedIn here:
https://www.linkedin.com/feed/update/urn:li:activity:7290455...
Comment · 1 year ago
Reminds me of:
https://distill.pub/2020/growing-ca/
"Differentiable Model of Morphogenesis"
- https://cautomaton.com
- https://jamandtea.studios
- https://earlgrey.ai
- https://bsky.app/profile/cautomaton.bsky.social