BoJangles

[BoJangles @yoshiTrain6]

@nikillinit The site is so slow bots gave up

[BoJangles @yoshiTrain6]

@pauliusztin_ @hwchase17 @LangChain https://t.co/Wuy6Mb2w4X

[BoJangles @yoshiTrain6]

@sebastiankehle I can’t even imagine the wasted tokens with this approach

[BoJangles @yoshiTrain6]

@alex_prompter But how well does it work in practice?

[BoJangles @yoshiTrain6]

@getpieces Wonder how you’d stack up on here: https://t.co/Wuy6Mb2w4X

[BoJangles @yoshiTrain6]

@eduardfausgil @leodev @supermemory This one uses a proxy approach. The model gets no say in whether it should fetch memory

https://t.co/LMLbGVIdLR

[BoJangles @yoshiTrain6]

@ptaranat @refactory What? I don’t think he was saying noreply is a subdomain tho

[BoJangles @yoshiTrain6]

@LightField @hwchase17 This works towards that; https://t.co/LMLbGVIdLR

[BoJangles @yoshiTrain6]

@billgnofficial Not me, but thought this was interesting; https://t.co/Wuy6Mb2w4X

[BoJangles @yoshiTrain6]

@LangChain @BraceSproul But how it performs is something to be desired

[BoJangles @yoshiTrain6]

@kunalstwt Vector search and AI memory do not belong in the same sentence

[BoJangles @yoshiTrain6]

@piyushsingariya But which ones actually work?

[BoJangles @yoshiTrain6]

@memco_ai @ScottTaylor You should get on the leaderboard: https://t.co/Wuy6Mb2w4X

[BoJangles @yoshiTrain6]

@somayyeh1990 There is already a benchmark that measures something different you are gesturing towards: https://t.co/Wuy6Mb2w4X

[BoJangles @yoshiTrain6]

@Jarvixlive Go ahead and get on the leaderboard

[BoJangles @yoshiTrain6]

@EzraDeveloper @leodev Nobody had is a bold claim

[BoJangles @yoshiTrain6]

@dreamd_dev Not entirely: https://t.co/8IypU2zC51

That one tests memory providers in isolation before any model

[BoJangles @yoshiTrain6]

@Ex0_Byte Where you at on precision though?

[BoJangles @yoshiTrain6]

@leodev Yeah looking real bleak