Andrew - building Performant

Currently working on Performant, a personal AI health coach on IOS.
You should try it out!

[Andrew - building Performant @andrewmcode]

@benhylak lol yes

[Andrew - building Performant @andrewmcode]

@ElliottMcNary @anothercohen Yea, bevel is good but AI features and nutrition aren't the best

[Andrew - building Performant @andrewmcode]

@mkarolian @lipbalmluvr @thatssovaibhav you should try Performant, refreshes take <1s
we do a history computation step which runs in the background and does take some time, but once that completes, refreshes are blazingly fast!

[Andrew - building Performant @andrewmcode]

@thatssovaibhav Pretty sure it's against whoop TOS if it bypasses the subscription

Apple watch is great, only downside is the battery life really

[Andrew - building Performant @andrewmcode]

@mal_shaik https://t.co/X5RDRWMuJT?

[Andrew - building Performant @andrewmcode]

@shubhdsgn Mx master 3, deathadder v3 pro

[Andrew - building Performant @andrewmcode]

@KrishnaBuildsAI Zed and Neovim are both better

[Andrew - building Performant @andrewmcode]

@daylenyang @kalo_nazih you can calculate strain and recovery without a whoop, if you have an apple watch you can get similar data without paying hundreds a year

[Andrew - building Performant @andrewmcode]

@sflorimm Performant

[Andrew - building Performant @andrewmcode]

@stylishdawg lol give him a return offer

[Andrew - building Performant @andrewmcode]

@vitaliidodonov MRR

[Andrew - building Performant @andrewmcode]

@OrionAlius @LexnLin most likely, this example is more realistic imo

[Andrew - building Performant @andrewmcode]

@robertcourson looks very nice! is there a way to hide the sidebar when expanded?

[Andrew - building Performant @andrewmcode]

@LexnLin what are you even making 🤯 144hrs is insane

[Andrew - building Performant @andrewmcode]

@kmdrfx what did I just watch

[Andrew - building Performant @andrewmcode]

Testing out a fully local version of Performant Chat, running Gemma 4-E2B. One thing that's concerning is moderate/high hallucination rate and tool call error rates.

Should we ship this as an option to Performant?

[Andrew - building Performant @andrewmcode]

@pontusab looks nice!
was there a reason to use 4.1 specifically? it doesn't do that well on visual reasoning benchmarks

[Andrew - building Performant @andrewmcode]

@TensorSlay m3 is surprisingly good, I wish it was faster though

[Andrew - building Performant @andrewmcode]

Just added dark mode, you can now view your health scores comfortably at night in Performant

[Andrew - building Performant @andrewmcode]

@teortaxesTex Doesn't anthropic / oai do the same?