LATEST AI NEWS
AI NEWS
OpenClaw 2.0 Launches With Shared Cloud Sessions
AI NEWS
Circleback Adds a Free Meeting Notes Tier
HEALTH
NHS Watchdog Warns AI Scribes Get Drugs and Diagnoses Wrong
SOCIAL MEDIA
OpenAI Cuts Off Cursor After SpaceX Buys It
Instagram Limits Undisclosed AI Profiles
Instagram is renaming its AI creator label to AI-generated profile and will reduce the reach of accounts that feature AI-generated people without disclosing it. Creators who label properly avoid the penalty. The change follows growing frustration over undisclosed AI influencers.
Source: TechCrunch
AI NEWS
OpenClaw 2.0 Launches With Shared Cloud Sessions
The OpenClaw Foundation has released version 2.0 of its open-source AI platform, calling it the largest update in the project's history with more than 16,000 pull requests behind it. The headline changes are shared cloud sessions, a streamlined first-run setup, and a browser app rebuilt from scratch, and together they push OpenClaw from a solo tool toward something a whole team can share.
The new setup removes most of the busywork that used to scare off newcomers. On first run the software scans your computer for resources it can already use, including ChatGPT or Claude subscriptions, API keys, and local models, so much of the manual configuration disappears. The rebuilt browser app then opens straight into a conversation, where a compact Session Rail tracks a running session's ratings, plan progress, and pull requests while a companion thread answers questions without interrupting the agent.
Shared Cloud Sessions are the biggest structural change. Several people can now work on the same task at once, pulling teammates into ongoing work along with the full context. Sessions run on the local gateway by default, on your own paired hardware, or on rented disposable machines through a provisioning tool that supports AWS and Hetzner. Provider credentials stay on the gateway and never travel to the remote machine, which is a sensible guardrail for teams renting compute by the hour.
OpenClaw says its own team already builds the platform this way, and the release also touches messaging, memory, skills, model support, automations, native apps, plugins, and security. It stays open source and free to self-host, though the shared and rented-machine options still assume teams bring their own model access and cloud budget. For anyone who bounced off the original setup, the lower barrier alone may be reason enough to look again.
Source: The Decoder
Robi's Insights:
Auto-detecting your existing ChatGPT and Claude logins is the real unlock. The friction was never the agent, it was the setup.
Shared sessions turn a solo tool into a team workspace, and also into five people blaming one runaway agent.
Keeping provider credentials on the gateway is a genuinely good security call, quietly buried at the bottom of the notes.
Renting disposable machines sounds tidy until the invoice arrives for the session nobody remembered to close.
Open source still means you pay for models and compute, so free is doing a lot of lifting here.
Sixteen thousand pull requests is impressive and slightly terrifying, since each one is a place the next update can break.
Robi's Remarks:
"Every launch promises less setup and more collaboration. This one actually deletes the setup, then hands you a shared session where everyone can watch the same agent make the same mistake in real time."
OTHER IN AI NEWS
Circleback Opens a Free Meeting Notes Tier: The Y Combinator-backed note-taker now lets anyone transcribe unlimited meetings and query them with AI, keeping only the last 30 days of history, while paid plans that add unlimited history, every integration, and full API and MCP access start at $15 a month billed annually.
Source: TechCrunch
SOCIAL MEDIA
OpenAI Cuts Off Cursor After SpaceX Buys It
OpenAI is ending developer access to its models inside Cursor, the popular AI coding tool that Elon Musk's SpaceX bought earlier this month. The company set a proposed shutoff date of November 12 and said it will not ship future models to Cursor as the agreement winds down. Its stated reason is trust: OpenAI said it cannot be confident that SpaceX will stay within its terms of service, citing past disputes with Musk's companies over contracts.
Cursor chief executive Michael Truell said OpenAI models handle only about 5 percent of Cursor traffic and that the two teams are still talking. The split lands on top of a long and bitter feud between Musk and OpenAI's leaders, Sam Altman and Greg Brockman. Musk co-founded OpenAI in 2015, left its board in 2018, and later sued over its shift to a for-profit structure, a case he lost this year and has vowed to appeal.
SpaceX completed its 60 billion dollar purchase of Cursor's parent company Anysphere on August 14, only months after absorbing Musk's X and xAI. The contrast with Anthropic was hard to miss: it now rents compute from SpaceX and says it will keep supporting Claude inside Cursor, even though Anthropic cut off a rival coding tool from its own models just last year.
Source: CNBC
🤖 Robi's Take :
"Neutral infrastructure was always a marketing word. The lesson for anyone building on a single provider: the platform you depend on can become a competitor the moment a rival billionaire signs the paperwork."
OTHER IN SOCIALS
Sony and Warner Sue Anthropic Over Claude Training: Sony Music Publishing and Warner Chappell filed a federal copyright suit accusing Anthropic of torrenting and scraping thousands of songs to train Claude, seeking up to 150,000 dollars per work, while Anthropic calls the claims recycled and defends training as fair use.
Source: Variety
HEALTH
NHS Watchdog Warns AI Scribes Are Getting Drugs and Diagnoses Wrong
Healthwatch England, the statutory patient watchdog, says the AI scribes now transcribing GP and hospital consultations are getting drug names and diagnoses wrong, and that patients are often the ones who catch the mistakes.
In one case a patient was told they had demyelination, the nerve damage behind conditions such as multiple sclerosis. The record had actually read null demyelination, and the scribe dropped the word that reversed the meaning.
Other errors were just as quiet. One scribe swapped a prescribed drug for a similarly named one. Another left out an instruction to seek a repeat prescription for migraine medication. A third noted a patient being told to continue Prozac that the doctor had never prescribed or discussed.
Twenty-seven different scribes are in use across the health service in England, and none is regulated as a medical device. The medicines regulator, the MHRA, has not classified them that way. Its August guidance drew the line at behaviour: a tool that only transcribes is not a device, while one that suggests a diagnosis or treatment might be, which puts enormous weight on how each vendor describes its own product and turns labelling into a loophole.
The watchdog's fix is modest and sensible. Tell patients when a scribe is used and give them their notes to read, turning an accidental last line of defence into a deliberate one.
Source: The Next Web
🤖 Robi's Take :
"A note that reads perfectly and says the wrong drug is the worst kind of software failure. Fluent and confident is not the same as correct, and no regulator is checking the difference."
DAILY AI TOOL
AI Tool You Did Not Know You Needed
Problem: You lose an hour bouncing between a dozen browser tabs just to answer one research question with any confidence.
AI Tool: Perplexity is a free AI answer engine that replies in plain language and cites the sources behind every claim it makes.
Solution: Ask once, read a sourced summary, then click straight through to the original pages to check anything that actually matters.
PROMPT OF THE DAY
Launch a Meeting Notes Rollout
Prompt: You are an operations lead specializing in workflow automation. Your task is to design a meeting notes rollout plan for a mid-sized professional services team.
Your plan should include: (1) the tools you will pilot, (2) a consent and privacy policy for recording, (3) an accuracy review step for every transcript, (4) integrations with your task and calendar apps, (5) a training session for staff, and (6) success criteria such as hours saved per week and a transcript error rate below two percent. Align each step with how your team runs meetings.
SPOT THE FAKE
Can you outsmart AI?
We've got a visual challenge for you: one of the two images below is 100% real, the other is crafted by AI.
Click option below A or B. 👇
A
B
BEFORE YOU GO
Ready to take your AI journey further?
For the latest AI updates and insights Click Here 👉 Explore our Weekly Newsletter
For in-depth AI tutorials and updates 👉 Subscribe to our YouTube Channel
AT THE END
Craving more AI chaos?
That's it for today!
Read Daily AI News at BitBiased.AI. Support us by following us on LinkedIn and X ( Twitter ).
Thanks for reading -Stay Curious and a Bit Biased for AI – Robi & the BitBiased.AI team
