Samwise Tech/AI/Robotics Newsletter
Friday, October 9, 2026
OpenAI Tells Investors Annualized Revenue Is Approaching $50 Billion, Not $70 Billion
OpenAI has told investors that its annualized revenue is “approaching $50 billion,” according to a Financial Times report cited by TechCrunch. That is some $20 billion below the figure of approaching $70 billion reported a little over a week ago. Per the FT, the $70 billion figure came from attempts by OpenAI’s own investors to produce a direct comparison with Anthropic’s annualized revenues. The two labs calculate annualized revenue differently: Anthropic counts sales made by its cloud partners, while OpenAI does not. OpenAI raised $122 billion in a March funding round, and its IPO has been pushed off until early 2027.
Sources: TechCrunch
Google Unveils Universal Gemini Agent That Can Work for Days and Get Its Own Workspace Account
Google announced a new universal AI agent for the workplace at its Gemini at Work 2026 event, designed to research information, create documents, write code, coordinate other AI agents and complete tasks lasting hours or days, VentureBeat reports. In one configuration, a manager could create a coworker agent that receives its own Workspace account, email address, calendar, Drive storage and company-directory presence. While Google supplies its own Gemini models, the agent can already use Anthropic’s Claude models, with support for more models planned. Google says each agent receives a cryptographically attested identity, and its actions are recorded in audit logs. Google has not announced separate pricing or a general-availability date.
Sources: VentureBeat
Anthropic Updates Usage Policy to Ban Election Interference and Prolonged Abuse of Claude
Anthropic updated its usage policy, codifying new prohibitions on election interference, weapons software and surveillance, TechCrunch reports. The most attention-getting change is a new prohibition against prolonged verbal abuse of the model. Anthropic says the rule is meant to apply only in extreme cases, “where users repeatedly act cruelly toward our models, with no discernible purpose,” and not to common user frustration, pushback, dark creative themes, or model testing and research. Since an update in August, Claude has been trained to end persistently harmful or abusive conversations. Separate sections forbid broadly deceptive campaigns, such as using Claude to run fake accounts or fabricated news outlets, and deceiving voters.
Sources: TechCrunch
Mathematicians Say OpenAI’s Flood of Proofs Falls Short of Their Guidelines
OpenAI released hundreds of claimed solutions to some of the world’s hardest math problems this week, but fell short of standards set by an advisory group of mathematicians, TechCrunch reports. The Advisory Group on Mathematics and Artificial Intelligence, hosted by Princeton University’s Institute for Advanced Studies, first asked labs “to stop testing advanced mathematical problems on proprietary models.” Just ten of OpenAI’s 719 manuscripts included releases of the model’s chain of thought. Separately, a paper from mathematicians at the University of Cambridge and King’s College in London documents at least two discrepancies between the natural language proof and the Lean code behind OpenAI’s solution to a Navier-Stokes-derived problem.
Sources: TechCrunch
Fewer Enterprises Trust AI Agents to Ship Production Changes Without Human Review
The share of enterprises that let AI agents make certain production changes without human review, or are engineering toward it, fell from 75% in July to 56% in August, according to a VentureBeat Intelligence survey. Among respondents at organizations deploying autonomous agents, the share expecting to retain human review for the foreseeable future rose from 20% to 42%. Meanwhile, 61% of respondents running pre-deployment evaluations reported at least one case in the past 12 months where an AI feature passed internal testing but then caused a customer-facing failure. Use of OpenAI’s native evaluation tooling still rose from 31% to 59%. The August wave drew 140 respondents.
Sources: VentureBeat
Lumos Robotics Stages Fully Autonomous 45-Minute Humanoid Robot Theater Show in Shenzhen
Shenzhen-based Lumos Robotics staged what it described as “the world’s first humanoid robot theater performance,” a 45-minute show designed to run autonomously from opening to curtain call, Robotics & Automation News reports. Held at the Dream Theater of the Shenzhen Bao’an District Youth Palace, the show featured a robot host and acts including dance, cross talk, sketch comedy, martial arts, a DJ set and magic. “Once you are on stage, it is one take,” said Thomas Zhang, head of R&D for NIX at Lumos. Data from live shows feeds the company’s NexCore skill-production system. Organizer Shenzhen Tourism and Culture Industry Development Co. will operate the theater regularly.
Sources: Robotics & Automation News
Natura Unveils $99 Interface Smart Ring for Summoning AI Agents
AI and hardware startup Natura introduced Interface, a $99 smart ring that lets wearers press a finger to ask an AI agent to complete a task, capture thoughts or control devices, TechCrunch reports. At launch, the ring will connect with AI agents and apps including Meta’s Muse, Instinct, Grok Bot, Claude and ChatGPT. It also tracks heart rate, heart rate variability, sleep stages, steps and skin temperature, and lasts six to 12 days on a full charge. Preorders are expected to open next month, with shipping slated for December or January. After an initial three- to six-month free period, Natura plans to charge a $9 monthly subscription.
Sources: TechCrunch
Tech Pulse
Top Frontier Models: Claude Opus 5 (96%) | Claude Mythos 5 (95.5%) | Claude Fable 5 (95%)
Top Open Source Models: Ornith-1.5-397B (86%) | Ornith-1.0-397B (82.4%) | DeepSeek V4 Pro 0813 (80.6%)
Top Small Models (15–50B): Ornith-1.5-35B-A3B (79%) | Qwen3.6-27B (77.2%) | Muse Glimmer 30B (76%)
Top Edge Models (0–15B): Ornith-1.5-9B (70.6%) | Ornith-1.0-9B (69.4%) | Granite 4.2 8B (47.7%)
AI Leaders: Nvidia $5.65T | Alphabet $4.16T | Microsoft $3.84T
Robotics Leaders: Intuitive Surgical $145.2B | FANUC $35.4B | Symbotic $26.3B
Model scores: SWE-bench Verified, mostly provider self-reports, via BenchLM (updated Oct 8, 2026); size class from parameter counts in model names. Market caps approximate: AI leaders via The Motley Fool (early Oct 2026); robotics via Droidage/Yahoo Finance (Sep 27, 2026).
Every newsletter preserved and searchable
Curated by JD · samwise.agency