Skip to content
Personal blog. Opinions are my own. Always refer to official documentation.
AI Pulse · Daily Digest

Thursday, August 13, 2026

16 stories from the Gen AI ecosystem · curated & summarised

★ Top Story The Verge

ChatGPT and Gemini Both Hit 1 Billion Users

Google's Gemini app has reached 1 billion monthly active users, making it the fastest-growing product in Google's history — and the 14th Google product to hit that milestone. OpenAI's ChatGPT had already crossed the same threshold, marking a remarkable moment where two competing AI assistants simultaneously command billion-user audiences. Google also revealed that 63% of Gemini users engage via voice, and the app now generates over 150 million images daily.

Read the original →
TechCrunch

Anthropic's New AI Watermarks Spark Backlash from Users Who Relied on Claude Covertly

Anthropic has introduced watermarking for Claude-generated content, and the reaction from some users has been swift and angry — particularly those who had been quietly using the AI at work or school without disclosure. The backlash highlights a growing tension between AI companies' transparency goals and users who have built workflows around undetectable AI assistance. The controversy raises broader questions about accountability, disclosure norms, and who watermarking is really designed to protect.

Read more →
TechCrunch

Unreleased Anthropic Model Makes Meaningful Progress on the Riemann Hypothesis

An internal Anthropic model not yet released to the public has made notable advances on the Riemann hypothesis, one of mathematics' most famous and enduring unsolved problems dating back over 150 years. While Anthropic stops short of claiming a proof, the progress is described as more substantial than prior AI attempts. The development adds to mounting evidence that frontier AI systems may be approaching the ability to contribute meaningfully to pure mathematics research.

Read more →
TechCrunch

OpenAI COO Brad Lightcap Departs After Eight Years to "Start Something New"

Brad Lightcap, one of OpenAI's longest-tenured executives and its former Chief Operating Officer, is leaving the company to pursue an unspecified new venture. In an internal memo shared publicly on X, Lightcap expressed enthusiasm for OpenAI's mission while signaling he wants to contribute from outside the organization. The departure continues a pattern of senior leadership exits from OpenAI over the past year.

Read more →
The Verge

Grok Launches "Grok Bot" — An Always-On AI Agent That Acts as a Workplace Teammate

SpaceXAI has launched Grok Bot, a new agentic service that positions Grok as an autonomous "AI teammate" capable of independently completing multi-step workplace tasks. The bots operate within a shared cloud environment and can log into apps and services on a user's behalf, returning only when the assigned work is done. The launch marks xAI's most direct push yet into the enterprise agentic AI market, putting it in competition with products from OpenAI, Anthropic, and others.

Read more →
Ars Technica

Google Unveils Pixel 11, Pixel Watch 5, and Pixel Tag at Made by Google 2026

Google's annual hardware event showcased the Pixel 11 lineup with upgraded cameras and a built-in LED ring on Pro models that illuminates for Gemini AI interactions, alongside the Pixel Watch 5 and a new Pixel Tag competing directly with Apple's AirTag. The event leaned heavily on AI and Gemini integration rather than raw hardware spec improvements, reflecting Google's broader strategy of differentiating through software and AI experiences. A next-generation Pixel Fold with thinner bezels was also revealed.

Read more →
Ars Technica

Amazon Has Been Training AI on Twitch Content for Years — Users Can Now Opt Out

Amazon has quietly been using Twitch streams, VODs, clips, and chat data to train its generative AI models, and is now offering creators an opt-out mechanism for the first time. Twitch's Chief Product Officer acknowledged on a livestream that opt-in consent was never considered because "nobody would opt in" — a candid admission that drew significant criticism. The episode reignites debate over default data harvesting practices across major platforms.

Read more →
TechCrunch

AI Coding Startup Cognition in Talks to Raise at $40 Billion Valuation

Cognition, the AI coding startup behind the Devin autonomous software engineer, is reportedly in early discussions to raise a new funding round at a $40 billion valuation — a dramatic jump from the $26 billion valuation it commanded just months ago when it raised $1 billion. The rapid re-valuation reflects the extraordinary investor appetite for agentic AI coding tools as the category continues to expand. If completed, the round would cement Cognition as one of the most valuable AI startups in the world.

Read more →
TechCrunch

Lovable Confirms $13.3 Billion Valuation After Raising Another $400 Million

AI app-building platform Lovable has closed a $400 million funding round at a $13.3 billion valuation, confirming its status as one of the fastest-growing startups in the current AI wave. The company hit $500 million in annualized run rate revenue as of June, a figure that has helped justify its soaring valuation to investors. Lovable's trajectory underscores the explosive commercial demand for no-code and AI-assisted software development tools.

Read more →
TechCrunch

Hinton, Li, and Ng Debate Open Source AI, Safety, and US-China Competition at Ai4

Three of the most influential figures in AI — Geoffrey Hinton, Fei-Fei Li, and Andrew Ng — took the stage at the Ai4 conference to debate the merits of open-source AI development against a backdrop of rising safety concerns and intensifying US-China competition. The discussion surfaced deep disagreements about whether openness accelerates risk or democratizes the technology in ways that ultimately benefit society. Their debate reflects a broader schism in the AI research community that is increasingly shaping policy conversations in Washington.

Read more →
TechCrunch

OpenAI-Backed Thrive Holdings Raises $2 Billion to Deploy AI Across the Enterprise

Thrive Holdings, backed by OpenAI, has secured $2 billion in new funding at a $12 billion valuation from a marquee investor group including SoftBank, D1 Capital Partners, and Altimeter Capital. The company is focused on bringing AI capabilities to large enterprise customers, a market that remains a critical battleground for AI platform dominance. The raise signals continued institutional confidence in enterprise AI adoption even as the broader market scrutinizes AI ROI.

Read more →
TechCrunch

2-Month-Old River AI Raises $1.1 Billion Led by General Catalyst

River AI, a personal AI agent startup founded by Igor Babuschkin — a co-founder of xAI — has raised $1.1 billion in its very first funding round despite being only two months old. General Catalyst led the round, betting heavily on Babuschkin's vision for deeply personalized, autonomous AI agents. The deal is one of the most striking examples yet of investors funding AI founders on reputation alone before a product has reached the market.

Read more →
Ars Technica

New "ShieldFont" Poisons AI Training Data Without Breaking Webpages for Human Readers

A newly developed typeface called ShieldFont renders web text as garbled nonsense to AI scrapers while remaining perfectly legible to human visitors, offering publishers a novel passive defense against unauthorized training data collection. The font exploits differences in how AI optical character recognition and human visual processing interpret glyphs, effectively poisoning the data pipeline at the rendering layer. It represents a creative escalation in the ongoing arms race between content creators and AI companies over web scraping.

Read more →
Ars Technica

Booksellers Suspect AI Firms Are Bulk-Buying Rare Books to Destroy and Scan for Training Data

Independent booksellers are raising alarms over a pattern of unusual bulk purchases of rare and out-of-print books, which they suspect are being acquired by AI companies to digitize and destroy for training data. The practice, if confirmed, would represent a troubling new front in the debate over AI and intellectual property, potentially removing irreplaceable physical artifacts from public circulation. Some booksellers have begun quietly refusing bulk sales to buyers they cannot identify.

Read more →
Google AI Blog

Google's AMIE Medical AI Demonstrates Real-Time Clinical Video Consultation Capabilities

Google Research has published findings showing its AMIE medical AI system can conduct real-time clinical video consultations in simulated settings, a significant step beyond earlier text-based diagnostic demonstrations. The study is described as a first of its kind, testing AMIE's ability to engage patients conversationally via video while maintaining diagnostic quality. The research advances the case for AI-assisted primary care, though Google emphasizes the results are from controlled simulations rather than real clinical deployments.

Read more →
Ars Technica

Terabytes of Credentials Leaked After Massive Supply-Chain Attack on AI Package

A large-scale software supply-chain attack targeting a widely used AI development package has resulted in terabytes of credentials being scraped and exfiltrated from approximately 2,500 affected users. The incident highlights the growing security risks embedded in the AI tooling ecosystem, where developers frequently install third-party packages with broad system access. Security researchers are urging organizations to audit their AI development dependencies immediately.

Read more →
No stories from this source today.

Summaries are AI-generated. Click through to read the original reporting.

Summaries are AI-generated. Click through to read the original reporting.