📍One Founder, Eleven Labs
Plus: Elevator Repairs Start With a Briefing, Bank Passwords Stay With You
Good Morning, Agent Owners!
Imagine being intelligent enough to accelerate drug discovery but getting fired for not logging in for 45 days.
One Founder, Eleven Labs
TL;DR: Mithril Medicine founder Cody Tranbarger used AI to coordinate contracts, logistics and experiments across 11 external labs, helping a one-person company move a potential muscular dystrophy treatment from idea to promising mouse results in under 18 months. Anthropic, which acquired Mithril, reports the progress. The research remains preclinical; human trials and an effective medicine are still ahead. Read more →
Drug Paperwork Loses 23 Hours
TL;DR: Enveda tested an AI system that checks pharmaceutical batch records against supporting documents and presents flagged issues to human reviewers. Across 1,061 pages from four records in one clinical manufacturing campaign, review time fell from 25.9 hours to under three. The company reports additional issues beyond those humans found. The 89% reduction concerns this review exercise, not all compliance work. Read more →
Elevator Repairs Start With a Briefing
TL;DR: Before entering a building, TK Elevator technician Torben Teichmann receives an AI briefing covering the equipment, fault and tools he needs. Afterward, a voice agent asks about the repair, requests missing details and records his answers for colleagues. Microsoft describes the workflow in use, without establishing a companywide rollout. The practical gain is less hunting for information and less typing. Read more →
Pichai Delegates the Daily Catch-Up
TL;DR: Google CEO Sundar Pichai says internal agents summarize model-training discussions and surface insights from compute usage, efficiency and utilization. He can then investigate selected issues himself. In an October 21 interview, he described work that could otherwise take an hour of daily reading. That is a personal estimate, but the workflow shows executives delegating routine monitoring while retaining judgment. Read more →
Amazon Pushes Back on Tokenmaxxing
TL;DR: Amazon’s updated engineering guidance tells developers to judge AI by lasting customer and business outcomes. It warns that targeting token consumption or code volume can inflate costs and complexity, while still allowing those metrics as carefully interpreted signals. Leaders must use the tools themselves. The memo makes understanding when AI helps, and when it does not, part of the job. Read more →
Telstra Checks the Migration Bill
TL;DR: Telstra is using an AI agent to compare billing and payment records across old and new systems during customer migrations. It flags inconsistencies, recommends corrections and prepares cases for human review. Microsoft says the production system supports more than 272,000 service migrations, with recommendation accuracy above 95%. That figure measures suggested actions, not the share of migrations completed without people. Read more →
Bank Passwords Stay With You
TL;DR: Bank of America is emailing customers a specific warning: do not give account usernames or passwords to outside services, including AI agents acting on your behalf. The notice also says customers remain responsible for activity resulting from shared access. As personal assistants take on financial errands, the bank is drawing an authorization boundary. The email does not establish any resulting fraud. Read more →
Trial Plans Get an Early Review
TL;DR: Bristol Myers Squibb is using AI to analyze proprietary data, develop clinical study plans and assess development options for an early-stage drug. Its early clinical development chief says an initial analysis would otherwise take days. The company and Anthropic plan broader work in 2027. This is an early planning application, with no demonstrated improvement in clinical trial success rates yet. Read more →
Truist Retires Idle Agents
TL;DR: Truist assigns an accountable owner to every AI agent and automatically disables agents after 45 days without activity. Microsoft says the bank manages more than 13,000 declarative agents, with over 8,000 active each month. The figures describe the supplier’s account of adoption and governance. The rule offers a concrete answer to a growing operational problem: abandoned agents retaining a place in company systems. Read more →
The Phone Queue Gets Delegated
TL;DR: A user says he asked an AI agent to contact a utility about account changes. It navigated automated phone menus, reached a human representative in roughly four minutes and dialed him into the call when the company needed to speak directly with him. The account illustrates a useful handoff, though the original post does not confirm the requested changes were completed. Read more →
Customers Bring Their Own Help
TL;DR: In a Capacity survey of 1,000 U.S. consumers, 35% said they had used a personal AI agent for a customer service request at least once. The share rose to 49% among both Gen Z and millennial respondents. Businesses may increasingly face customers assisted by their own agents. These are vendor-sponsored, self-reported findings, rather than measured adoption or successful resolution rates. Read more →
Asylum Summaries Face Scrutiny
TL;DR: A UK immigration watchdog is considering scrutiny of AI tools that summarize asylum interviews and retrieve policy information, following Open Rights Group research questioning accuracy and transparency. The systems began pilots in late 2024. No inspectorate investigation has been confirmed. The Home Office says trained caseworkers make decisions, but inaccurate summaries could still shape the evidence those people rely on. Read more →
Your agent cloud, everywhere — in your chat apps, browser, and tools. Always on.
- Add to Slack, Teams, Telegram or WhatsApp.
- With a clean, redesigned WebUI.
- Connect any of 100+ apps & sources in one click.
- GPT 6 Astra, DeepSeek v4.1 & more frontier AI models, built in at discounted rates.
Run your agent on MyClaw.ai in Seconds👇
