AJ Chan (@itsajchan) 's Twitter Profile
AJ Chan

@itsajchan

Growth @Weaviate_io. Building in public and having fun with Gen AI, RAG, vector databases, 3d printing and more

ID: 1345391906

linkhttps://ajchan.io calendar_today11-04-2013 22:02:10

3,3K Tweet

2,2K Followers

968 Following

AJ Chan (@itsajchan) 's Twitter Profile Photo

We running Calibrate Hack Day right now! Tons of prizes, incredible questions from the community, and fantastic workshops from Weaviate โ€ข vector database, Comet, CrewAI, LlamaIndex ๐Ÿฆ™. Tons of support from other incredible companies too!

AJ Chan (@itsajchan) 's Twitter Profile Photo

Had several failed deliveries from another vendor which forced me to do overnight prints FedEx. Thanks for making it happen! Now I get to run Calibrate from my living room with my doggo, Morty

Had several failed deliveries from another vendor which forced me to do overnight prints <a href="/FedEx/">FedEx</a>. Thanks for making it happen!

Now I get to run Calibrate from my living room with my doggo, Morty
AJ Chan (@itsajchan) 's Twitter Profile Photo

Hack Night GitHub on June 18th is going to be dope. We have one more spot left for a developer tool that fits the stack. Anyone interested in joining?

Hack Night <a href="/github/">GitHub</a> on June 18th is going to be dope. We have one more spot left for a developer tool that fits the stack. Anyone interested in joining?
Perly (@_shark_byte) 's Twitter Profile Photo

Another hackathon win for the week๐Ÿฅณ Used Weaviate โ€ข vector database to make a cli tool that recognizes when you have sensitive info copied to your clipboard (like API keys, or passwords etcโ€ฆ) Thanks so much to Weaviate โ€ข vector database Microsoft , AJ Chan and Comet for the plushie ๐Ÿ˜Š

AJ Chan (@itsajchan) 's Twitter Profile Photo

๐Ÿ’กYour document search is probably lying to you. It says your complex PDFs are searchable. It says your apps can โ€˜understandโ€™ content. But the moment you throw in mixed layouts, embedded charts, or a random screenshot, infused documentsโ€ฆ everything breaks. Well, youโ€™re not

๐Ÿ’กYour document search is probably lying to you.

It says your complex PDFs are searchable. It says your apps can โ€˜understandโ€™ content.

But the moment you throw in mixed layouts, embedded charts, or a random screenshot, infused documentsโ€ฆ everything breaks.

Well, youโ€™re not
AJ Chan (@itsajchan) 's Twitter Profile Photo

๐—ฌ๐—ผ๐˜‚๐—ฟ ๐—”๐—œ ๐—ฎ๐˜€๐˜€๐—ถ๐˜€๐˜๐—ฎ๐—ป๐˜ ๐—ถ๐˜€ ๐˜๐—ฟ๐—ฎ๐—ฝ๐—ฝ๐—ฒ๐—ฑ ๐—ถ๐—ป ๐—ฎ ๐—ฑ๐—ฎ๐˜๐—ฎ ๐—ฝ๐—ฟ๐—ถ๐˜€๐—ผ๐—ป. Here's how MCP is breaking down the walls: Anthropic unveiled Model Context Protocol (MCP), an open-source standard designed to bridge AI models directly with the data stored in business apps,

๐—ฌ๐—ผ๐˜‚๐—ฟ ๐—”๐—œ ๐—ฎ๐˜€๐˜€๐—ถ๐˜€๐˜๐—ฎ๐—ป๐˜ ๐—ถ๐˜€ ๐˜๐—ฟ๐—ฎ๐—ฝ๐—ฝ๐—ฒ๐—ฑ ๐—ถ๐—ป ๐—ฎ ๐—ฑ๐—ฎ๐˜๐—ฎ ๐—ฝ๐—ฟ๐—ถ๐˜€๐—ผ๐—ป.

Here's how MCP is breaking down the walls:

Anthropic unveiled Model Context Protocol (MCP), an open-source standard designed to bridge AI models directly with the data stored in business apps,
Sammy Sidhu (@sammy_sidhu) 's Twitter Profile Photo

Today we're announcing that Eventual has raised $30M in Seed and Series A funding from CRV and Felicis as well as Y Combinator, M12 - Microsoft's Venture Fund and Citi and others. The AI era needs data infrastructure built for AI, not retrofitted. ๐Ÿงต

AJ Chan (@itsajchan) 's Twitter Profile Photo

If you're using an LLM to answer a prompt... ๐˜ต๐˜ฉ๐˜ข๐˜ตโ€™๐˜ด ๐˜ˆ๐˜. ๐—œ๐—ณ ๐—ถ๐˜ ๐—ฟ๐—ฒ๐—ฎ๐˜€๐—ผ๐—ป๐˜€, ๐—ฝ๐—น๐—ฎ๐—ป๐˜€, ๐˜‚๐˜€๐—ฒ๐˜€ ๐˜๐—ผ๐—ผ๐—น๐˜€, ๐—ฎ๐—ป๐—ฑ ๐—ฎ๐—ฑ๐—ท๐˜‚๐˜€๐˜๐˜€ ๐—ฏ๐—ฎ๐˜€๐—ฒ๐—ฑ ๐—ผ๐—ป ๐—ณ๐—ฒ๐—ฒ๐—ฑ๐—ฏ๐—ฎ๐—ฐ๐—ธ...Thatโ€™s an agent in action. The distinction between static AI models and dynamic agents is becoming

If you're using an LLM to answer a prompt... ๐˜ต๐˜ฉ๐˜ข๐˜ตโ€™๐˜ด ๐˜ˆ๐˜.

๐—œ๐—ณ ๐—ถ๐˜ ๐—ฟ๐—ฒ๐—ฎ๐˜€๐—ผ๐—ป๐˜€, ๐—ฝ๐—น๐—ฎ๐—ป๐˜€, ๐˜‚๐˜€๐—ฒ๐˜€ ๐˜๐—ผ๐—ผ๐—น๐˜€, ๐—ฎ๐—ป๐—ฑ ๐—ฎ๐—ฑ๐—ท๐˜‚๐˜€๐˜๐˜€ ๐—ฏ๐—ฎ๐˜€๐—ฒ๐—ฑ ๐—ผ๐—ป ๐—ณ๐—ฒ๐—ฒ๐—ฑ๐—ฏ๐—ฎ๐—ฐ๐—ธ...Thatโ€™s an agent in action.

The distinction between static AI models and dynamic agents is becoming
AJ Chan (@itsajchan) 's Twitter Profile Photo

๐—ง๐—ฟ๐—ฎ๐—ฑ๐—ถ๐˜๐—ถ๐—ผ๐—ป๐—ฎ๐—น ๐—ฒ๐—บ๐—ฏ๐—ฒ๐—ฑ๐—ฑ๐—ถ๐—ป๐—ด๐˜€ ๐—ด๐—ผ๐˜ ๐—ถ๐˜ ๐—ฎ๐—น๐—น ๐˜„๐—ฟ๐—ผ๐—ป๐—ด. Semantic search isn't about squeezing meaning into a single vector - it's about preserving every nuance. The problem with single-vector embeddings no one discusses: Converting "A very nice cat" into one