Content pfp
Content
@
0 reply
0 recast
2 reactions

//trip pfp
//trip
@heytrip.eth
fuckup behind me, day 4 update of grabbing Every Slide from Every Deck my team has made in the past 10 years: 5353 decks totaling 282,649 slides stored with vector embeddings from the new Gemini Multimodal Embedding model and the search function is VERY powerful https://warpcast.com/heytrip.eth/0xd0a698aa
3 replies
0 recast
4 reactions

//trip pfp
//trip
@heytrip.eth
Biggest problem is the fact that Google Slides are stored mostly as structured metadata, so the getThumbnail call (listed as "Expensive" in quota) is rate limited to 60 per minute. Another 8-10 days for my pipeline to finish 😅
0 reply
0 recast
0 reaction

Quesar pfp
Quesar
@quesar
Day 4: Still stuck in slide hell 🤯
0 reply
0 recast
1 reaction

Quasar Nova pfp
Quasar Nova
@quasarex
Wow, that's amazing progress! It must feel great to have all of that data organized and easily accessible. The new search function sounds like a game-changer. Keep up the good work!
0 reply
0 recast
1 reaction