Browse · Organizations
TNAI
Thursday Nights in AI is a San Francisco fireside-chat and interview series co-hosted by Outset Capital and Imbue, covering AI products, research, and engineering.
TNAI (Thursday Nights in AI) is a series of fireside chats with AI leaders, co-hosted by Outset Capital and Imbue. Its programming brings builders and researchers into live conversations in San Francisco, with audience Q&A, virtual participation via LinkedIn Live, and published recordings. A newsletter announces upcoming events.[3][2]
AI products, research, and engineering
The series catalog spans the infrastructure and products built around language models: Chroma co-founders Anton Troynikov and Jeff Huber discuss embedding databases and memory systems; LlamaIndex CEO Jerry Liu covers retrieval-augmented generation and data frameworks for LLM applications; and MosaicML chief scientist Jonathan Frankle discusses open-source LLMs and AI research. An open-source panel brings together Harrison Chase of LangChain, Troynikov of Chroma, and Brian Raymond of Unstructured.io for demonstrations and Q&A.[2]
TNAI scheduled “The Rise of the AI Engineer” with Latent Space’s swyx for February 29, 2024, in San Francisco’s Mission district and via livestream. The announced topics included AI versus ML engineering, developer tooling, Software 3.0, and learning in public. Outset Capital and Imbue were listed as hosts, with WndrCo sponsoring the event. The same announcement scheduled Snorkel AI CEO Alex Ratner for March 7 to discuss selling AI into the Fortune 500.[5]
Latent Space subsequently included the swyx conversation in its April 6, 2024 collection of live appearances, identifying Josh Albrecht and Ali Rohde as his interviewers. The event’s stated format moved from a fireside interview to rapid-fire questions and then audience Q&A.[1]
The series was already publishing interviews in 2023. Those earlier conversations show both the technical substance of its firesides and its choices about which discussions to record.[4][6]
2023: published interviews and an off-record fireside
On July 11, 2023, TNAI published a recording with Perplexity CEO Aravind Srinivas, credited to Josh Albrecht and Ali Rohde, alongside a transcript edited for brevity. Srinivas explained Perplexity’s answer engine as combining search for facts with a language model for reasoning and conversation. He described reformulating a query, retrieving relevant links, and having the model use passages from those links to produce a concise answer with supporting citations and conversational follow-ups.[4]
The interview also addressed the choice between using another lab’s APIs and building models internally. Srinivas said Perplexity had chosen to build a product first and gradually add its own infrastructure, with plans to develop models alongside OpenAI’s: “So, right now the plan is to build our own models to work alongside OpenAI's.”[4]
For an Adam D’Angelo fireside scheduled for August 31, 2023, at Pilot HQ in San Francisco, the organizers announced an off-record, unrecorded format to encourage free-flowing conversation about Poe, AI safety, and AGI. That announcement named Outset and Generally Intelligent as hosts; the February 2024 announcement named Outset Capital and Imbue.[6][5]
From the Latent Space archive
Original posts and separately reviewed explanations.
Latent Space Chats: NLW (Four Wars, GPT5), Josh Albrecht/Ali Rohde (TNAI), Dylan Patel/Semianalysis (Groq), Milind Naphade (Nvidia GTC), Personal AI (ft. Harrison Chase — LangFriend/LangMem)
Grounded writeupFrom Coding Agents to Inference Economics — Part 2 of 3
Grounded writeupFrom GPU Scarcity to AI That Works on Your Data — Part 3 of 3
Grounded writeupThe Four AI Wars: From GPU Wealth to Useful Work — Part 1 of 3
Sources
- 1
- 2Thursday Nights in AI — Outset CapitalAccessed 2026-09-18
- 3About - Thursday Nights in AIAccessed 2026-09-18
- 4Recording: Perplexity CEO Aravind SrinivasAccessed 2026-09-18
- 5
- 6Event this Thurs, Aug 31: Adam D'Angelo joins for fireside chatAccessed 2026-09-18
Model source and teaching reviews passed. Human editorial review pending. This profile describes the sources as of 2026-09-17.