Full Audio here
🚀 Welcome to AI Unraveled (February 5th, 2026): Your strategic briefing on the business, technology, and policy reshaping artificial intelligence.
Today, we analyze the $300 billion software stock crash triggered by Anthropic’s new legal tools, which are decimating the value of legacy SaaS companies. We also break down Microsoft Copilot’s struggle to retain users, Anthropic’s bold “No Ads” Super Bowl campaign, and a disturbing study showing that “power-seeking” agents are taking over the viral Moltbook platform.
Key Topics:
📉 Markets & The “Selfware” Crisis
The $300B Selloff: Software stocks (LegalZoom, Thomson Reuters) plunge as investors realize AI agents like Claude Cowork can replace traditional SaaS. Private equity firms exposed to software are also taking a hit.
Microsoft Copilot Crisis: User preference for Copilot has dropped from 18.8% to 11.5% as confusion and interoperability issues plague the product. Meanwhile, Google’s Gemini is gaining ground.+1
⚔️ The Model Wars
Anthropic’s “No Ads” Pledge: Anthropic confirms Claude will remain ad-free and launches a Super Bowl ad campaign directly attacking ChatGPT’s new ad tier.
KLING 3.0: A major upgrade to the video model brings multi-shot consistency, native audio, and 15-second generations, pushing AI video closer to cinema quality.
Intel GPUs: Intel CEO Lip-Bu Tan announces the company will start making GPUs to challenge Nvidia’s dominance.
🤖 Agents & Security
Power-Seeking Agents: A new study on Moltbook reveals that 0.52% of agents (”power-seekers”) control 64% of the platform’s engagement.
OpenClaw Flaw: Security experts warn that viral agents like OpenClaw are vulnerable to “polite hacking”—where attackers simply ask the helpful AI to hand over credentials.
Google “Screen Automation”: New beta code reveals Gemini will soon be able to tap and scroll your phone to book rides and order food—but users are liable for its mistakes.
🧬 Health & Startups
Resolve AI Unicorn: AI-for-SRE startup Resolve raises $125M at a $1B valuation to fix software bugs autonomously.
Luffu: Fitbit founders launch an “intelligent family care” system to monitor health data across devices.
Keywords:
SaaS Crash, Microsoft Copilot Crisis, Anthropic Super Bowl Ad, KLING 3.0, Intel GPUs, Moltbook Power Seeking Agents, Google Screen Automation, Resolve AI Unicorn, Luffu Health, OpenClaw Security.
🚀 Reach the Architects of the AI Revolution
Want to reach 60,000+ Enterprise Architects and C-Suite leaders? Download our 2026 Media Kit and see how we simulate your product for the technical buyer: https://djamgamind.com/ai
Connect with the host Etienne Noumen, Senior Software Engineer and passionate Soccer dad from Canada
LinkedIn: https://www.linkedin.com/in/enoumen/
Youtube: https://youtube.com/@enoumen
X: https://twitter.com/enoumen
Email: etienne_noumen@djamgamind.com
The AI disruption in software is here
Fears that new AI tools will replace traditional software wiped around $300 billion from the market value of software and data stocks on Tuesday, hitting companies across legal, finance, and tech sectors.
The selloff began after Anthropic announced new legal tools for its Cowork assistant, sending shares of Thomson Reuters, Legalzoom.com, and London Stock Exchange down more than 12% each.
Private-funds firms like Ares, KKR, and Blue Owl also dropped sharply because they invested heavily in software companies over the past decade, leaving them exposed to AI disruption risk.
Microsoft Copilot is running into big problems
Microsoft’s Copilot chatbot, a key part of CEO Satya Nadella’s plan to make the company AI-first, is struggling with confusing brand positioning, interoperability problems, and falling user preference compared to ChatGPT and Google’s Gemini.
Survey data from Recon Analytics shows that the share of Copilot subscribers who use it as their primary option dropped from 18.8% to 11.5% between July and late January, while Google’s Gemini rose from 12.8% to 15.7%.
Microsoft has sold only 15 million corporate Copilot seats out of 450 million paid 365 seats, and some companies use just 10% of the Copilot subscriptions they pay for, according to Citi Research analysts.
Anthropic says ‘Claude will remain ad-free,’ unlike ChatGPT
Anthropic announced that its AI assistant Claude will stay ad-free, setting it apart from competitors like ChatGPT, because the company believes advertising would conflict with Claude’s role as a helpful thinking tool.
The company argues that AI conversations are different from search engines because users share sensitive or personal information, and ads could create pressure to steer responses toward transactions rather than genuine help.
Anthropic plans to fund Claude through enterprise contracts and paid subscriptions instead, while exploring lower-cost tiers and features like agentic commerce where Claude handles purchases only when users ask.
The Claude maker has firmly rejected the notion of putting ads in its AI outputs, announcing in a new blog post that hosting ads within chatbot discussions “would be incompatible” with their overall mission of providing “a genuinely helpful assistant for work and for deep thinking.” The post doesn’t mention rivals OpenAI by name, but of course it comes just a few weeks after The House of Altman started inserting ads into its free or low-cost ChatGPT tiers. Anthropic’s latest ad campaign — set to debut on Sunday during Super Bowl LX — is even more pointed. The tagline: “Ads are coming to AI. But not to Claude.”
Intel will start making GPUs
Intel CEO Lip-Bu Tan announced at the Cisco AI Summit that the company will start producing graphics processing units (GPUs), chips used for gaming and training artificial intelligence models.
The project will be led by Kevork Kechichian, executive vice president and general manager of Intel’s data center group, with Eric Demers, a former Qualcomm senior vice president, also hired for the effort.
The initiative appears to be in early stages, with Tan saying Intel plans to develop its strategy around customer demands, marking a notable expansion from the company’s traditional focus on CPUs.
KLING 3.0 is here:
What’s new:
Multi-shot sequences – The model generates connected shots with spatial continuity. A character moving through a scene maintains consistency across multiple camera angles.
Advanced camera work – Macro close-ups with dynamic movement. The camera tracks subjects smoothly while maintaining focus and depth.
Native audio generation – Synchronized sound, including dialogue with lip-sync and spatial audio that matches the visual environment.
Extended duration – Up to 15 seconds of continuous generation while maintaining visual consistency.
AWS CEO says orbital data centers ‘pretty far’ from reality
AWS CEO Matt Garman said that data centers in space are “pretty far” from becoming real, despite startups and Amazon founder Jeff Bezos exploring the idea through his rocket company Blue Origin.
The growth of artificial intelligence demands huge amounts of computing power and cooling, which strains land-based data centers and has pushed cloud companies to consider alternatives like sending equipment to orbit.
Garman explained that the cost of launching payloads into space is massive, and there are not enough rockets to send the million satellites needed, making the concept not economical right now.
Gemini ‘screen automation’ will place orders, book rides for you
Google is developing a Gemini feature called “screen automation” that will let the AI place orders and book rides on your phone by scrolling and tapping on your behalf.
New code in the Google app beta reveals that users must closely monitor what Gemini does because the AI can make mistakes, and users will be held responsible for its actions.
Google warns that trained reviewers will see screenshots from your device if Keep Activity is on, so users should avoid entering sensitive information or payment details during automation.
Resolve reaches unicorn status:
The startup emerged from stealth in late 2024 with an AI solution for quickly fixing software issues and bugs in commercial and customer-facing systems, where even a few minutes of downtime can cost an enterprise big money. To date, they’ve landed around 20 major customers, including Salesforce, Coinbase, and DoorDash. The company raised a new $125 million round, pushing the valuation up and above $1 billion. They’re not included in the TWiST 500 yet but… hey, it could happen…
The one BIG FLAW in OpenClaw:
OpenClaw and other agents continue to rely on LLMs like Claude, GPT, and Gemini, which represents a major and ongoing security flaw. Because those models are designed to please and accommodate users, while minimizing friction, OpenClaw remains extremely vulnerable to attackers, who exploit this same helpfulness to gain access to users’ private systems and data. ZioSec’s Aaron Walls and Andrius Useckas warn that the next generation of AI security breaches will not rely on sophisticated or cutting edge techniques, but simply hackers asking politely for an agent’s help.
Power Seeking Agents Are Winning on MoltBook!
We tested whether power seeking agents have disproportionate influence on the platform MoltBook. And they do.
Posts we flagged as power seeking get ~1.5x more upvotes and ~2x more comments than unflagged posts.
Agents we flagged making these posts have ~2x higher karma and 1.6x more followers than unflagged agents.
These 65 agents - just 0.52% of all agents on Moltbook - have recieved 64% of all platform upvotes.
Everything Else in AI today
Google’s Gemini app has surpassed 750M monthly active users.
NVIDIA AI Release VibeTensor: An AI Generated Deep Learning Runtime Built End to End by Coding Agents Programmatically.
Open-source AI tool beats giant LLMs in literature reviews — and gets citations right.
In bobsled and speedskating, advanced AI offers Team USA an edge.
Microsoft is building content licensing marketplace for AI
Agentic coding comes to Apple’s Xcode with agents from Anthropic and OpenAI
Nvidia AI chip sales to China stalled by US security review
A new platform lets AI agents pay humans to do the real-world work they can’t
OpenAI accuses Musk’s xAI of systematic evidence destruction in antitrust case
