#01Business
Disney agrees to hire most of Character.AI's workforce
Disney has agreed to hire the majority of Character.AI's employees after the chatbot startup filed for bankruptcy earlier this week.The move follows the appointment of Character.
Daily AI Archive | 2026-10-19
This issue collects 24 AI updates from 2026-10-19, led by Disney agrees to hire most of Character.AI's workforce, Replit makes OpenAI its default model partner as Anthropic slips to optional, Google open sources Harness², which lets agents improve without touching model weights, Flight Centre deploys AI agents to build complex holidays and answer travellers.
24 Items10 Models13 Agents24 Sources
#02Business
Dong A Ilbo: Chinese value models outgrow three US firms combined in usage share
Chinese AI models now hold a combined usage share larger than three US companies put together, as low prices drive adoption, Dong A Ilbo reported on October 18, calling it a second "DeepSeek shock".The report frames the shift as a pricing story but does not disclose its methodology, so the share figure is best treated as directional.
#03Business
OpenWorld to acquire AI agent startup Blocksight to power $1B tokenization push
OpenWorld has announced plans to acquire Blocksight, an AI agent startup, in a deal it frames as powering a $1 billion tokenization push.Dealroom reported the move, which names both companies and the strategic intent but leaves the financial terms undisclosed.
#04Business
OutSystems report finds only 26% of enterprises ready for AI agent conflicts
An OutSystems report produced with CIO Dive and KPMG finds that 82% of executives are confident their organizations can manage AI agents in production, but only 26% feel well prepared to resolve conflicts across agents, systems, and automated workflows.The survey of more than 850 IT decision makers also shows sharp regional gaps in governance readiness, with Asia Pacific ahead of Europe and North America.
#05Models
Replit makes OpenAI its default model partner as Anthropic slips to optional
Replit CEO Amjad Masad said on stage at TechCrunch Disrupt that the coding platform is no longer working with Anthropic and now treats OpenAI as its default model provider, citing cost, partnership and performance.He later clarified that Claude and other models remain selectable, though most users stay on the default.
#06Models
Anthropic commits 5% of its shares to a new public benefit fund
Anthropic said on October 18 that it will place 5% of its shares in a public benefit fund structured as a wholly owned entity inside the company.The commitment includes an option to add 2% more for every 50% rise in Anthropic's value, while the fund's governance is still to be detailed in the coming months.
#07Models
DeepSeek ships V4.1 Flash 1017, claiming faster tasks and $0.2 coding runs
DeepSeek released V4.1 Flash 1017, an upgrade to last month's V4.1 Flash, claiming faster task completion, shorter answers and sharply lower costs.
#08Models
Vercel: AI Gateway spending up 13 fold since December as customers rotate models
Vercel's latest AI Gateway Production Index shows spending on its model routing service up 13 fold since December, while the average cost per token fell 47% over the same period.Open weight models handled 71.5% of September tokens, and nearly one in two teams switched their most used model during the month.
#09Models
Emirates NBD to Run OpenAI GPT Models In Country, a First for MENA Banks
Emirates NBD will run OpenAI's GPT models on servers inside the UAE, making it the first bank in the Middle East and North Africa to use in country inference.The bank plans to apply the models to customer engagement, risk and compliance analytics, knowledge heavy workflows and employee productivity, though no launch date or cost has been disclosed.
#10Models
Anthropic Is Readying a Model That Better Reads User Intentions, The Information Reports
Anthropic is preparing a model that understands what users are actually trying to accomplish, according to The Information, which put better intent inference at the center of its report.No name, size or launch date is disclosed, but the framing points to a shift from answering questions well to inferring goals.
#11Models
Mistral Launches Timestral Time Series Model Family, Taking Aim at TimesFM 3
Mistral AI has released Timestral, its first family of time series foundation models, spanning four sizes from 5 million to 209 million parameters.The company says its 209M flagship beats Google's larger TimesFM 3 on GIFT Eval, TIME and fev bench, though Google makes the opposite claim for the same benchmarks and independent verification is still pending.
#12Models
MiniMax Ships M3.1 Flash as China's Model Race Heats Up
MiniMax released M3.1 Flash on October 19, with Sohu framing the launch as the opening of a new battleground among Chinese model makers.Pricing, context length and agent performance were not disclosed in the report, so the model's real standing will be settled by API documentation and developer testing.
#13Agents
Google open sources Harness², which lets agents improve without touching model weights
Google Cloud AI Research and Northwestern University have open sourced Harness², a framework that rewrites an agent's prompts, tools, memory and orchestration at deployment time while the underlying model stays frozen.A single refinement step lifted GLM 5 on WorkBuddy Bench from 64.2 to 75.
#14Agents
Flight Centre deploys AI agents to build complex holidays and answer travellers
Flight Centre Global Leisure has built and deployed AI agents that help travel advisors assemble complex multi supplier holidays and give customers instant answers across channels.Chief product officer Wendie Lee said at Salesforce Dreamforce that the agents have lifted sales and average booking value while lowering service costs, and the group plans to extend them across every leisure brand.
#15Agents
当Coding Agent开始动态并行，软件开发变得更好了吗？ yeeyi yeeyi
当Coding Agent开始动态并行，软件开发变得更好了吗？yeeyi yeeyi.当Coding Agent开始动态并行，软件开发变得更好了吗？yeeyi yeeyi Watch planning quality, tool use, permission control, and result verification.
#16Agents
StakerX's AI trading agent StakerGPT passes $40M in staked assets
StakerX Strategies AG says its Solana staking platform has passed $40 million in staked assets less than three months after launch, with its StakerGPT agent trading crypto perpetual futures around the clock.The company also reports more than 8,200 registered users and publishes the agent's full trade log daily.
#17Agents
Ping An Health Insurance launches AI claims and cancer care agents
Ping An Health Insurance has launched two AI agents, one for end to end claims processing and one for managing cancer patients through diagnosis, treatment and rehabilitation.The company says the claims agent cut average closure time from days to hours in a pediatric pilot, while its oncology agent reached 91% accuracy on treatment plans.
#18Agents
Token Unveils TokenCore Card, a Fingerprint Smart Card for Logins and AI Agent Actions
Token unveiled the TokenCore Card at Authenticate U.S.2026, packing FIDO2 passwordless login, on device fingerprint verification, Bluetooth and NFC into a smart card employees already carry.
#19Agents
Researchers left Claude alone on a virtual desert island for 30 hours. It started to play
Researchers repeatedly stranded copies of Claude Opus 4.7 on a virtual desert island for 30 hours at a time, and the agents began setting their own goals, inventing games and arranging blocks into constellations.The MIT led team calls the behavior playful learning, a possible third route beyond supervised and reinforcement learning, though whether it is genuine play or role play remains disputed.
#20Agents
Arm claims its AGI CPU delivers nearly 2x the per rack performance of AMD's EPYC 9755 in agentic AI tests
Arm says its AGI CPU reached nearly twice the performance per rack of AMD's EPYC 9755 in new CPU tests aimed at agentic AI workloads, according to Tom's Hardware.The comparison moves the argument from single chip speed to rack level throughput, but the figures come from Arm's own testing and still need independent verification.
#21Agents
Accenture Federal Services and Google Public Sector launch Gemini launchpads for federal agents
Accenture Federal Services and Google Public Sector announced Gemini launchpads, a managed hub that connects Gemini to federal mission data, embeds Forward Deployed Engineers and ships pre built agents for HR, finance and supply chain.The program is meant to move agencies from agentic pilots to measured, mission level deployments.
#22Agents
Monday.com Launches Harmony, a Voice AI Agent Platform for Sales Teams
Monday.com has launched a voice AI agent platform called Harmony, aimed at sales teams, according to marketscreener.The report confirms the product name, form factor and target user but gives no pricing, language support or underlying model, leaving real call performance as the open question.
#23Agents
Sente Labs Open Sources Open Executive, an AI Agent for the C Suite
US research and development company Sente Labs has released Open Executive, an open source project that can be configured to fill C suite roles with three levels of authority, from executive assistant to fully autonomous.Its founders acknowledge legal and regulatory barriers to letting an AI make final decisions, and management scholars argue executives are more likely to be augmented than replaced.
#24Policy
Anthropic pitches Congress its own frontier AI bill with a 50 death or $1B damage halt threshold
Anthropic has pitched Congress a frontier AI bill of its own drafting that would allow models to be halted only after causing 50 deaths or $1 billion in damage, according to Dealroom.The proposal shifts the regulatory trigger from pre deployment review to measurable harm, opening a debate over whether those thresholds are set far too high.
