@mrdrozdov
Search and Agents @ Databricks
AI
Databricks says the architecture combines episodic and derived memory through plugins, on-demand tools and offline consolidation.
AI
A post calls them the “best company” and “best city” but does not name either.
AI
A researcher introducing the benchmark says the top model scored 16.6%, while meeting physical requirements remains a major challenge.
AI
A post describes “science again” as a common refrain at COLM, while questioning whether RL loss tweaks are better.
AI
A keynote listener highlights the idea that stronger models will need longer prompts because people will demand more of them.
AI
A user quote-posted the Recurse Center link with a fire emoji and no further explanation.
AI
One of the paper's authors reports more diverse, less repetitive creative output with XTC, a sampling method that can exclude high-probability choices for what comes next.
AI
A September 25 announcement says the work has an oral slot in the conference’s E&D track, celebrating recognition for research focused on users.
AI
Turing Post’s September 22 roundup includes JEPA-Anything, DeepSeek-V4.1-Flash and several papers on recursive self-improvement, alongside an AliceAI model release.
AI
A user expressed excitement while quoting another post.
AI
Creator Diogo Almeida claims Jev is 20–200x faster and 40–400x cheaper, with free output tokens. Investor DCVC says it led TypeSafe’s $40 million seed round.
AI
A post sharing a Weaviate Podcast episode says the key lesson was giving an analysis agent tools to investigate why relevant documents ranked poorly, rather than relying on a score alone.
AI
“Bricked my Jarvis,” a user wrote, describing the tension between big Codex projects and keeping daily email automations running within usage limits.
AI
Perplexity founders confirm appointment with focus on open research in continual learning and multi-agent systems.
AI
The release treats ColBERT-style late interaction models as a first-class model type for training, inference, and interpretation.
AI
Four former Google researchers found a public benefit corporation to automate machine learning and science.
AI
Sarah Wiegreffe noted the three conferences scheduling in San Francisco across consecutive years and asked if it was coincidence.