Michael

xLabs | on TG is not me

[Michael @mcyliu]

@ibuprod @lolo_IRL @saucepoint 💀💀💀

[Michael @mcyliu]

@tmuxvim waiting for this to go viral

[Michael @mcyliu]

@henryksarat i’m still waiting for the city to cut a tree overhanging the house for the past year…lol

[Michael @mcyliu]

@ibuprod smoggy

[Michael @mcyliu]

@ibuprod @lolo_IRL @AnthropicAI bro it kicked me off for asking “continue”

[Michael @mcyliu]

@BenedettoBio distribution distribution distribution distribution

[Michael @mcyliu]

@ibuprod sup hottie

[Michael @mcyliu]

@ibuprod @lolo_IRL you’re crushing it man!!

[Michael @mcyliu]

@henryksarat real intents imo

[Michael @mcyliu]

@ibuprod @lolo_IRL this is awesome man. i wish you all of the success Ibu! great showcase of real use case!

[Michael @mcyliu]

@ibuprod i see drinking

[Michael @mcyliu]

@uosvald @Yieldy_io awesome urban!! looking fwd to it!

[Michael @mcyliu]

@tmuxvim 😭🤣🤣

[Michael @mcyliu]

@ibuprod @lolo_IRL @worldbuildlabs @worldnetwork @FWBtweets @Consort_Tech @worldcoinfnd let’s go my dude!!

[Michael @mcyliu]

@ibuprod @iBUKAR3L @sama if that’s the case, i guess my question is, what’s the advantage of using openclaw other than ease of use to message it via tg/wa? CC is already the orchestration agent and can write its own prompts based on the context of your codebase, tho multimodel vs codex/cc separately…

[Michael @mcyliu]

@ibuprod @iBUKAR3L @sama ya for sure. if you have checked out the GSD skill on CC, def check it out. the built up memory is rly just a ton of .md files anyways for openclaw, also opus is 1M token context now too. my general opinion is no need for openclaw for the actual coding but open to other opinions

[Michael @mcyliu]

@ibuprod @iBUKAR3L @sama yea hard to say with the limited knowledge i have of your workflow, but presumably, the agent is just taking your QA inputs and then scaffolding plan -> solution -> sprint > test. memory and session constraints are “easier” to get around now since you can have CC write memory

[Michael @mcyliu]

@tmuxvim 🤣🤣