
Building @inferablehq
Adelaide, Australia.
Joined March 2013
- Tweets155
- Following287
- Followers149
- Likes335
John Smith retweeted
We unbundled and released our structured results provider as a standalone project.
Introducing L1M (pronounced el-one-em) 🚀
John Smith retweeted
By popular demand, we're also launching pgchat as a single executable!
npx pgchat postgresql://user:pass@host:5432/db
👆 is all you need to chat with your postgres.
John Smith retweeted
When a human-facing system becomes sufficiently complex, "vibes" become a perfectly valid evaluation methodology
John Smith retweeted
We've just refreshed our console to feature the adapters! Let @trpcio, @apollographql, @PostgreSQL or @ValDotTown be the tooling runtime for your agents.
John Smith retweeted
229 ⭐️ Thank you! 🙏
star-history.com/#inferableh…
github.com/inferablehq/infer…
#starhistory #GitHub #OpenSource via @StarHistoryHQ
John Smith retweeted
Replying to @marcklingen @swyx
An OSS maintainer who has built an integration with Langfuse for @inferablehq -- love to see how everyone can "focus on what makes their beer taste better" by working on integrations
John Smith retweeted
its has come to my attention that many of you aren't actually sure what an ai actually agent is
So here is a little diagram explaining an agent (simplest form) vs api call + tool
John Smith retweeted
Build conversational LLM applications with Inferable and assistant-ui inferable.ai/blog/posts/assi…
John Smith retweeted
A secure bridge between your data systems and Inferable, enabling natural language queries while keeping your data private inferable.ai/blog/posts/anno…
John Smith retweeted
→ @inferablehq is a managed developer platform designed to simplify the creation of secure and reliable AI applications powered by LLMs.
It provides a managed solution for developers who want to focus on building their applications without managing complex LLM infrastructure.
🔴 Live on DevHunt → devhunt.org/tool/inferable
John Smith retweeted
We just dropped a blog post about doing on-device local RAG using Ollama, sqlite, sqlite-vec and mxbai-embed-large.
John Smith retweeted
👉 Something we found interesting about "needle threading" in LLMs.
🎯 LLMs can perform inference on inputs up to their context limits, their effective context limit is much shorter and their performance decreases as the context window grows.
John Smith retweeted
Never seen this alert before – Apple now tells you if an app you're deleting has a subscription still active. Good move.
(Taken on iOS 13 beta 2.)
I appreciate that @AppleMusic has a lyrics section. Probably not necessary for dance tracks though. – at Malvern, Adelaide
1
8





















