reality_inspctr

No agents.md?

dig the idea.

love this. I did something similar w the US Constitution. https://usconstitutionapi.com/

this is cool, I used to work in ag robotics and understand the economics require bravery, and loads of project finance eventually?

one thing: the fish cursor on the site is frustrating pls allow disabling.

This is a great setup. I was building agents in 2023 with email addresses, and it was very effective for onboarding tech laggards to generative AI. It also resulted in very low durable use rates. Users would onboard then stop using quickly. Christenson ftw again.

One thing that can be handled much better today than back then was MIME / attachment types and having agents assemble "fancy" email structures irt.

https://github.com/realityinspector/ATAT

pemmican is pretty easy to make if you modify to a ground texture, grind the meat, and accept that it's not "traditional" and refrigerate it

nice work. dealt with same issue during helene. would be interesting to do things like convert to morse code or convert to modem-over-walkie low baud rate.

I built this repo as a Helene response repo, trying to use an llm to help get resources over text message. https://github.com/realityinspector/supply_drop_ai

wonder if you could get to news over sms, use an llm to compress to minimum viable text?

amazing

who's gonna train a model on these?

I would be interested in this as well, but would love it to be more like openapi protocol, with an agents.md.

yeah fastapi + htmx is very effective.

excited to try out. flask <> coding agents work really well together.

cool idea will check it out

very cool

[yawn] at the meta narrative comments. let's just have fun:

1) Canada and Mexico. The inevitable rise of the US will erode borders where languages are shared. Mexican-American tech workers will pass advantage to Spanish-speaking friends and relatives. Canadians will host maple syrup breakfast meetings with American innovators from Toronto.

2) The Bahamas for obvious reasons

3) Extremely cold and extremely hot countries where it's miserable to be outside part of the year; Matrix-style AI+VR headsets will offer relief. Aka the sun lamp holodeck theory.

Martial _arts_

Somewhat related: here's an open source llm expense labeler for small businesses. https://github.com/realityinspector/llm_expense_labeler/ [OC]

I tried to build a startup based on a similar idea, but the traction was not good. The squarespace website is still up https://sundialcalendar.com/

It’s stunning how much of our lives are predictable. We are free will surfers on waves of determinism.

this is great! very clever logic model.

I wrote something I think is highly relevant about a distributed L1 for time, that you might enjoy, OP?

[oc] https://www.seanmcdonald.xyz/p/the-clockchain-protocol-the-l...

Hi OP - you might be interested in something I tried to launch but didn't succeed at getting traction. sundialcalendar.com

The idea is a calendar for your whole life. Would love to find a new home for it.

synths https://twitter.com/seanmcdonaldxyz/status/16611400526747443...

ey HN,

Introducing synthasAIzer, a developer-centric tool harnessing the power of AI to generate synthetic characters in expansive and dynamic networks. It's not just your standard character generator – it's designed to provide detailed personas and complex relationship structures. It's available at https://patch-synthasaizer.rltyinspctr.repl.co/.

This version demo creates synths and branches networks; we're going to be rolling out module after module, just like an analog synthesizer. Master plan here: https://strange-loop-synthasaizer.replit.app/

Using it is as easy as tweaking a few parameters. Choose the depth of relationships, the number of personas per role, specific character traits, and boom! The AI goes to work. The output? Comprehensive JSON files, ready to be plugged into whatever narrative, game, or simulation you're creating.

And here's where it really shines – we've tuned the AI specifically for character generation. This isn't a vanilla model. We feed it rich character-related data, pushing the AI to avoid generic or repetitive characters and relationships. The characters it creates are unique, dynamic, and tailored to your specifications.

For example, if you request a network of characters set in a medieval realm, synthasAIzer not only provides the individual personas, but also their connections and to some extent, dynamics within the network (https://patch-synthasaizer.rltyinspctr.repl.co/).

We designed synthasAIzer with developers in mind. You could be creating the next big RPG, a dynamic narrative, or just playing around with AI-driven character design - synthasAIzer is built to deliver.

This is just the start, and we're eager to develop synthasAIzer further. Would love to hear what you think, HN.

it can be helpful, for some, to give up on the notion of retirement altogether.

exist to exist the whole way through. do what you gotta do.

many SV people have probably also seen anecdotal evidence the $B crowd has this mindset. it's the $M crowd sweatin the small stuff.

"We are releasing AlpacaFarm, a simulator enabling everyone to run and study the full RLHF pipeline at a fraction of the time (<24h) and cost (<$200) w/ LLM-simulated annotators. Starting w/ Alpaca, we show RLHF gives big 10+% winrate gains vs davinci003 (http://crfm.stanford.edu/2023/05/22/alpaca-farm.html)"

Wendy used to hand-draw album covers for tapes of tests she sent to Bob Moog when he sent prototypes to try.

They are beautiful and childlike with wonder. A couple went to Smithsonian.

source: I’m former cochair of Moog Foundation

good.

that would be the goal, but it's hard. I went down a long blog-post road about a "clockchain" which includes a way to reconcile zone-less time. tl;dr all time is relative between a minimum of two objects. all but certain there is an elegant way to do it beyond my current thinking.

https://www.seanmcdonald.xyz/p/the-clockchain-protocol-the-l...

cal.com is solving some of these problems with handling of all zones.

Super cool. Applied to keep building a calendar that lets you Spend Time from a time budget.