Model Releases

Hosting the 2nd founder dinner in sf on August 31 with @GuangyuRobert at @Fundamental - the team behind @tryshortcutai Agents are increasing…

Hosting the 2nd founder dinner in sf on August 31 with @GuangyuRobert at @Fundamental - the team behind @tryshortcutai Agents are increasingly capable of solving tasks for longer task horizons without

DGX agentx-post
model-releasesjerry-liu--x

Hosting the 2nd founder dinner in sf on August 31 with @GuangyuRobert at @Fundamental - the team behind @tryshortcutai Agents are increasingly capable of solving tasks for longer task horizons without human in the loop. This has rewarded "horizontal" products owned by the frontier lab, like Codex and Claude Code, which let users specify a broad range of tasks through a single interface. For startups to survive, is the moat around: - Product (can you build a differentiated interface from Cowork) - Harness (can you build a more differentiated agent for a given set of tasks) - Data (can you accumulate a proprietary set of data that frontier models are not trained on) - Deep domain and workflow expertise (can you learn more about a business than OpenAI can) - Brand / mindshare (you're more well known for a particular thing than a frontier model) RSVP here: https://luma.com/llamai-8hry

Related

Source: Jerry Liu (X) | 2026-08-26

Loading related sources…