Tee Ploy Gen

Learning AI in public from every walk of life.

[Tee Ploy Gen @teeploygen]

@Ronny_MiniMax Please add the capability to enabled/disables tools in mcode and minimax code.

I don't use media generation much and it's injected on every turn.

20k token overhead on every turn is not nice.

[Tee Ploy Gen @teeploygen]

@pidotdev why is pi so slow at updating new models? i wanna try out glm-5.3-flash!

[Tee Ploy Gen @teeploygen]

@opencode How is it the same usage like Minimax M3 when glm 5.3 flash is cheaper by 2x

[Tee Ploy Gen @teeploygen]

@AiHubMix @Zai_org @ZaiforStartups Wow! Cheaper and with vision. Sign me up!

[Tee Ploy Gen @teeploygen]

@Alibaba_Qwen Better and cheaper than DSV4F, count me in!

[Tee Ploy Gen @teeploygen]

@Da7_Tech By glm according to Bloomberg

[Bloomberg @business]

The new AI model that’s swept to the top of online usage charts with high performance at zero cost was created by China’s https://t.co/q5KG9bMPxw, also known as Zhipu https://t.co/mYlzIWdtqQ

[Tee Ploy Gen @teeploygen]

@MiniMax_AI Wen new model? M3 is legacy model at this point

[Tee Ploy Gen @teeploygen]

Well, that was fast. Here's what I did with qwen3.8 max:
- Audited codebase & refactored it
- Added new feature to my personal tradingview indicator
- Created a pi extension
- Review and update bunch of .md files and plans
- Generated a few images using qwen 3.0 image

7-day usage limit burned in a few hours. Renew never!

[Tee Ploy Gen @teeploygen]

Today's goal: Burn all my usage with Qwen3.8 max which will expire soon.

Qwen token plan limits are not great with 3.8 max and it's very slow.

[Tee Ploy Gen @teeploygen]

Today's goal: Burn all my usage with Qwen3.8 max which will expire soon.

Qwen token plan limits are not great with 3.8 max and it's very slow.

[Tee Ploy Gen @teeploygen]

@Da7_Tech Is there a site to view how models perform according to your benchmark?

[Tee Ploy Gen @teeploygen]

@FradSer @pidotdev Starred! Will check it out later.

[Tee Ploy Gen @teeploygen]

@pidotdev Thank you for this! My agent keep eating $ when using bash. This fixes it!

[Tee Ploy Gen @teeploygen]

@MiniMax_AI @SambaNovaAI Nobody cares. M3 is so bad at following instructions.

[Tee Ploy Gen @teeploygen]

@RenLeanna When minimax 3.1?

[Tee Ploy Gen @teeploygen]

@ZixuanLi_ Won't use it until it has vision

[Tee Ploy Gen @teeploygen]

@jun_song About time!

[Tee Ploy Gen @teeploygen]

@mattpocockuk @theo Thank you for this! Finally, it's cleared to me which one I need to use.

[Tee Ploy Gen @teeploygen]

Finally, did it!

My own simple windows workspace that allows me to work with one pi agent that can delegate work to fresh new pi sessions. Inspiration drawn from First Mate by @kunchenguid

Goodbye pi-subagents, welcome observability!🥳

[Tee Ploy Gen @teeploygen]

@taohe669360 can't actually DM @MiniMaxAgent because their inbox is closed but yes, i want to be part of the beta

[Tee Ploy Gen @teeploygen]

@_overment how is this possible without using subagents?