R to @0xPaulius: 👋 🕶️

R to @Luc_AI_Insights: @AirOpsHQ is doing some cool stuff with agents + looking for talent

R to @MannyBernabe: Looking for a no code platform that creates calendars

Missed Kyle Dempsey's talk on building with LLMs? Watch the replay now at https://youtu.be/t4IHBo0BZz0 and learn how @airopshq is revolutionizing the way teams build and scale AI applications. #AI #LLMs

R to @sethjkramer: You could do this in @AirOpsHQ in like half the steps and half the time 🚀

R to @MannyBernabe: Hey Manny, You seem to know what you are talking about, is there anything available that fulfills these criteria? (I am not well versed) - Let me connect to multiple API's, get output and keep iterating, like comfy UI for Stable diffusion, but for everything - Ideally node based

🔧 The LLM Low-Code Tool Stack - Bot building: @getbotpress , @VoiceflowHQ - Custom LLM Services: @StackAI_HQ , @FlowiseAI - Automation & Integration: @zapier , @pipedream - AI Workflows: @AirOpsHQ

R to @bentossell: Here are some of my favs: @StackAI_HQ -- Easy to use and integrates well with other systems. @getbotpress & @VoiceflowHQ -- Bot builders with strong native LLM features. @AirOpsHQ -- Wide breadth of use cases, from chatbots to batch operations. @FlowiseAI -- Free open-source langchain-ish no-code GUI. Dark Horse: @zapier -- A lot of LLMs need an automation to be helpful. Zapier is building out some really neat "micro-apps" they call "interfaces". Haven't tried it, but it looks like a great play. Right now, I mostly use @pipedream for automation, but it would be nice to have easier UIs to tap into.

R to @alexhalliday: Yo Alex is GPT-4V confirmed for API?

Huge news this morning. We spend hours every day working with LLMs to solve a variety of generative use cases. @AnthropicAI's models are a key part of our playbook, accounting for 40% of the 1m LLM calls we've seen in the last few weeks at @AirOpsHQ. Their models are performant, uniquely large in context window and reliable. Current TPMs are 2x the closest comparable model from other providers. Excited to see what they ship next 🚀

R to @alexhalliday: very excited to use!

Huge day for image generation models. Finally we'll have a model that can actually incorporate text in it's output. We'll be adding to @AirOpsHQ as soon as it becomes available. New use cases incoming!

Now available in all AirOps Workspaces on all plans

.@OpenAI just released GPT-3.5-turbo-instruct 🎉 Using chat-tuned models in instances where you need simple instruction adherence is 👎 gpt-3.5-turbo-instruct will be live on AirOps early tomorrow, promising for classification, text extraction and other text analytics use cases

R to @valtterikaresto: The diversity of applications for this stack mean there will be a diversity of UI layers. Pairing (co-pilot) vs supervised agents vs fully automated systems all present new challenges. A lot to figure out as a category. / AH

R to @swyx: I Agree. I think UI layer won't be just chat nor traditional interface. We have some sort of transition happening. So we have started to call it Intentface, transition from interfaces to intentfaces.

R to @astralwave: just like the system of record => system of reasoning change, UI layer probably wont look like UI layer

R to @swyx: UI layer would be something like @AirOpsHQ (B2B) or Poe (B2C), no?

R to @dnlbtlr: 4. Middleware / app builders e.g. @FixieAI @retool @bentomlai @AirOpsHQ @humanloop @StackAI_HQ @scale_AI @ReworkdAI 5. Vector databases e.g. @zilliz_universe @lancedb @pinecone 6.Model Development, hosting, deployment: e.g. @huggingface @Skymel_ai @replicatehq ...