Tim Seckinger

Software Engineer / Engineering Manager, mostly on the web.
Tech Lead for the @framer CMS

[Tim Seckinger @_jeysal_]

Here's my conclusion after 1.7 seconds (253 US dollars) of coding with GPT 5.6 Sol:

[Tim Seckinger @_jeysal_]

once again, core engineering skills *always come back* - this time it's testing

[Tim Seckinger @_jeysal_]

i didn't think i would enjoy ai harness engineering this much before we built one @framer

[Tim Seckinger @_jeysal_]

and for false positives, because of determinism, you'd probably not hit them until someone refactors something a year on and you notice the test assertion was overly specific, and then you'd go relax the test once needed. makes refactors annoying but not that big a deal

[Framer @framer]

We’re live. Tune in now as we count down. https://t.co/9XZN9mxLdS

[Jurre Houtkamp @jurrehoutkamp]

June 16 @framer

[Christoph Nakazawa @cnakazawa]

@_pedro__soares There is no performance difference between Jest and Vitest. The issue is either your config, or most likely, the code you are testing.

[Addy Osmani @addyosmani]

https://t.co/ze3x7bgsL4

[Jacob 'kurtextrem' Groß @kurtextrem]

https://t.co/24Vb506gzz

[kitze @thekitze]

gpt 5.5 might be agi

[Tim Seckinger @_jeysal_]

@tilljanek @jonastreub @_CalMorris @framer https://t.co/lV8ejc9D8U

[Tim Seckinger @_jeysal_]

@tilljanek @jonastreub @_CalMorris @framer Let's just say we're not throwing overboard *everything* ;)

[Tim Seckinger @_jeysal_]

ln -s AGENTS.md README.md