←Search results

Search “Agents”

By relevance · Total 6704 results Clear search

Wednesday

  2. Huxiu

    Instinct founder on an app-less personal Agent: How a 14-person team supports a $10 billion valuation

    Instinct, an AI personal agent company founded a year ago with 14 employees, has reached a $10 billion valuation. Founder Noah Shinn discussed its product and business model on the podcast "Invest Like The Best." Instinct has no standalone app; users book flights and cancel subscriptions through text messages, phone calls, and emails, and its annualized transaction volume has reached the $1 billion scale, about half of which comes from business travel. It is free for users, and the company charges commissions to merchants that profit from transactions; Noah explicitly rejects the advertising model. About 40% of users link a credit card after roughly three weeks of use, and this group has a retention rate of about 80%.

    Instinct, an AI personal agent company founded a year ago with 14 employees, has reached a $10 billion valuation. Founder Noah Shinn discussed its product and business model on the podcast "Invest Like The Best." Instinct has no standalone app; users book flights and cancel subscriptions through text messages, phone calls, and emails, and its annualized transaction volume has reached the $1 billion scale, about half of which comes from business travel. It is free for users, and the company charg

    Matched fieldsTitle、Body

    #hot-signal#Industry news#Agents#Speech
  6. Tencent Cloud Developer Community

    TencentDB Agent Memory Adapts to the JEV Decision Model: Letting Agents Call on Team Experience as Needed

    TencentDB Agent Memory introduces a pluggable decision service and adapts to the JEV decision model, delegating "whether to recall, which candidate assets to select, and whether to trigger a memory update" to an independent interface for judgment. In Skill recall tests, the input tokens repeatedly injected by the main model dropped from 24.96M to 15.04M, a reduction of nearly 40%, while the task success rate remained essentially flat; the number of long, difficult problems passed on SWE-bench Pro rose from 30% to 40%. On the memory generation side, total cost fell by 13.8%, L3 generation calls decreased by 39.7%, and JEV judgment cost accounted for only 0.72% of total cost.

    TencentDB Agent Memory introduces a pluggable decision service and adapts to the JEV decision model, delegating "whether to recall, which candidate assets to select, and whether to trigger a memory update" to an independent interface for judgment. In Skill recall tests, the input tokens repeatedly injected by the main model dropped from 24.96M to 15.04M, a reduction of nearly 40%, while the task success rate remained essentially flat; the number of long, difficult problems passed on SWE-bench Pr

    Matched fieldsTitle、Summary、Body

    #hot-signal#Product updates#Agents#RAG

Tuesday

Monday