Thomas Dijkstra

Building & marketing real-life experiences for 200k+ visitors/year, 8-figures revenue.

Goal: bring 1M people together / year.

Posts on, AI, code, ads, ops.

[Thomas Dijkstra @thomasdijkstra]

@brendt_gd Did you replace phpstan/others with mago or do you use it in conjunction?

[Thomas Dijkstra @thomasdijkstra]

No drama in my style checks 💁🏼‍♂️ #codex

[Thomas Dijkstra @thomasdijkstra]

ChatGPT Pro since ‘24😎

Thanks for the gift @OpenAI @sama @gdb

Cool idea, definitely feel special now🙃

[Thomas Dijkstra @thomasdijkstra]

@zeeg i think this is the general consensus, at least here on x

[Thomas Dijkstra @thomasdijkstra]

@Kappaemme1926 yes feels like it lost some common sense but could be me

[Thomas Dijkstra @thomasdijkstra]

@mbertulli for more than a decade

[Thomas Dijkstra @thomasdijkstra]

@SpainProgrammer @simonhamp this ^ symfony is severely underrated

[Thomas Dijkstra @thomasdijkstra]

@norbert_tech Currently using phpstan, rector and csfixer. Saw mago pop up a while ago but it wasn't really there yet.

Is it as good as phpstan etc nowadays? Or still missing a lot of rules/features?

[Thomas Dijkstra @thomasdijkstra]

@ChereneAubert would love this :)

[Thomas Dijkstra @thomasdijkstra]

@ZssBecker quality content

[Thomas Dijkstra @thomasdijkstra]

@levelsio my wheels on my classic rimowa are also quite shitty and stuck and the handle wobbles 🤷‍♂️.

[Thomas Dijkstra @thomasdijkstra]

Claude be like: "so this is grounded, not hand-wavy."

[Thomas Dijkstra @thomasdijkstra]

queued 321 agents with the new claude code workflow feature to improve every integration test file in my codebase.

[Thomas Dijkstra @thomasdijkstra]

anyone having this with opus 4.8?

[Thomas Dijkstra @thomasdijkstra]

hmm

[Jarred Sumner @jarredsumner]

Dynamic workflows and adversarial code review was part of what made it possible to rewrite Bun in Rust in 6 days.

[Thomas Dijkstra @thomasdijkstra]

casually doesn’t mention the actual cursor bench score 👀?

[Cursor @cursor_ai]

Claude Opus 4.8 is now available in Cursor.

On CursorBench, it's able to work much more efficiently than Opus 4.7. We've also found it to be more persistent on harder tasks.

[Thomas Dijkstra @thomasdijkstra]

rip opus 4.8

[Thomas Dijkstra @thomasdijkstra]

@itsolelehmann bro it already dropped

[Thomas Dijkstra @thomasdijkstra]

69,2% for SWE-bench-pro is wild curious to try opus 4.8

[Lisan al Gaib @scaling01]

Claude Opus 4.8 Benchmarks