Will Bird

Cybersecurity & Automation Engineer

[Will Bird @wb1936]

Step 1: Use Pi, proven to cost the least
Step 2: Use my Aptus skill or just Caveman/Ponytail Ultra, if you actually like reading code the agent made do NOT do this
Step 3: Use a codex advisor skill and luna max as the implementation agent

All skills are here

https://t.co/ShLfddLyZ8

[Will Bird @wb1936]

@pilvar222 As an independent researcher? O think that is my main problem

[Will Bird @wb1936]

@pilvar222 Nice work

[Will Bird @wb1936]

@m9rcn poor pricing when https://t.co/hgna9T9Lhe exists

[Will Bird @wb1936]

I built Asset Curator for @NousResearch’s Hermes Agent hackathon.

Ask for a game asset: find an existing file or create one from code, with licences and provenance kept.

29 source channels, 18 built-in generators, 16 MCP tools. No extra model required.

Demo below 👇

[Will Bird @wb1936]

I made a cute poster for the sticker packs I got from @OpenAI DevDay London! Let me know if anyone else has their sticker packs and wants this file.

[Will Bird @wb1936]

@avec Just you wait until there is a AR based giveaway and I will win it...
6 13 6 92

[Will Bird @wb1936]

It's unfortunate people cannot "look" at concepts. Like symbolically/conceptually look at a bow only one man can string. It isn't because of some physical reason or some science, it's because that's conceptually how it is within the epic. The world's strongest man today couldn't do it, because that's just how it is.

[SB 🪖 @welovesweetboy_]

what was the actual reason odysseus was the only one capable of stringing his bow? was it because he was the only one worthy, or was it simply a matter of physical strength?

[Will Bird @wb1936]

@TheRealSorcus Yea I would appreciate that

[Will Bird @wb1936]

@TheRealSorcus Any specific or less known optimizations you are running for this or?

[Will Bird @wb1936]

@avec Any chance you email put people's score or placement?

[Will Bird @wb1936]

Code based generations will be the future as it's just more accessible for agents to pick up and use on the spot.

I don't know what you guys are working on internally, but personally I'd recommend making code based generation for all asset types; vfx, sprites, skeletons, animations, sounds etc and I might do some research into this stuff.

[Will Bird @wb1936]

I like to use it to run light benchmarks and it's stalled on those and also on one's where I was asking Astra to do weight-space programming.

When it happens I will just no longer get updates from that session and I have to ask a new chat for the progress on it, if it's still working etc.

[Will Bird @wb1936]

@feedthejim Web

[Will Bird @wb1936]

@feedthejim ChatGPT Work on wen isn't stable enough and my sessions often stall. Let me know if you want the session IDs

[Will Bird @wb1936]

@paw_lean The second I hit 500 LinkedIn connections I'm done

[Theo - t3.gg @theo]

The $200 Claude sub is the best money you can spend as an engineer right now. Truly unbelievable value.

I don't care how "cheap" other models are. I don't care that you can host your own inference with shitty models on a $10,000 GPU.

There is no place you can put $200 and get more value out the other side.

[Will Bird @wb1936]

@elonmusk @bot I promise separate usage pools is the way