Today's AI Pulse
Pinterest inks $4 billion AWS cloud pact to power AI visual search
Pinterest announced a $4 billion multi‑year agreement with Amazon Web Services, the largest infrastructure deal in its history, to power AI‑driven visual search and discovery through 2031. The contract expands use of AWS Trainium chips for large‑language and vision‑language models and adds a migration to a Kubernetes‑based architecture.
Also developing:
A Mom and Tech Entrepreneur Building AI Advocate for Rare-Disease Families Like Hers
Citizen Health, founded by tech entrepreneurs Farid Vij and Nasha Fitter, has launched an AI‑powered "AI advocate" to help rare‑disease families manage appointments, insurance appeals, and clinical‑trial connections. The platform, backed by $44 million in venture capital—including a $30 million Series A—now supports over 8,000 patients across 350+ conditions and partners with more than 16 pharmaceutical firms. By aggregating consented medical and genetic data, the service offers free, real‑time assistance while feeding researchers a high‑quality dataset that can shave 30‑50% off drug‑development timelines. The AI advocate aims to return the average 53 hours per week parents spend on care back to families and accelerate cures.
Paid to Film Tasks for AI, Ethics Ignored
there are businesses that pay people to do tasks while videoing themselves to train AI to do the same tasks later using robots. Nobody in this entire pipeline is questioning if we should even be doing it in the first...
How OpenAI's Codex Figured Out How to Use Adobe Software
OpenAI's Codex was used to batch‑denoise 50 images in Adobe Lightroom without any official API or plugin. Engineer Peter Gostev prompted Codex to discover a way to automate the task, and the model successfully interfaced directly with the desktop application....
"Catch Me If You Can": DT Using AI to Kill Deepfakes
Telecom voice networks are being overrun by fraud, with more than half of global calls now deepfakes, scams or extortion attempts, and the rate exceeds 60% in Mexico. This crisis has driven users to ignore unknown calls and rely on...
AI Momentum Shifts Toward Anthropic and Other Breakouts
At a major gathering of tech leaders, the spotlight was on Anthropic, with strong interest in its coding tools and growing market presence. The takeaway is simple. Momentum in AI can shift quickly, and right now, attention is clustering around a...
AI-Supported Scans Measuring Heart Fat Could Better Predict Cardiovascular Risk
Researchers used artificial intelligence to automatically quantify pericardial fat on routine coronary artery calcium (CAC) scans in a cohort of nearly 12,000 adults followed for about 16 years. The AI‑derived fat volume was independently associated with higher long‑term cardiovascular disease...
Google's Gemma 4 Puts Free Agentic AI on Your Phone and No Data Ever Leaves the Device
Google unveiled Gemma 4, an open‑source, on‑device AI model that handles text, images, and audio without sending data to the cloud. The model ships in four sizes—E2B and E4B for smartphones, 26B and 31B for servers—and is bundled with the free...
SoundHound AI Teams with Associated Carrier Group to Bring Agentic AI to Tier‑2/3 Telecom Operators
SoundHound AI has signed a partnership with Associated Carrier Group (ACG) to provide its agentic AI platform to ACG’s Tier 2 and Tier 3 telecom members. The deal gives smaller carriers access to autonomous AI agents that can resolve customer issues end‑to‑end,...
Oracle’s AI Push Triggers Record Credit Risk and 30,000 Job Cuts
Oracle announced a reduction of 30,000 jobs while its credit default swap spread hit a record 198 basis points, reflecting investor alarm over $124.7 B of non‑current debt tied to aggressive AI spending. The company must convert a $553 B backlog into...
Berkshire Hathaway HomeServices Rolls Out AI‑Powered BOSS Platform for 2,500 Agents
Berkshire Hathaway HomeServices The Preferred & Stouffer Realty launched AI BOSS, an AI‑infused evolution of its Business Operating Success System, to equip its 2,500+ agents with practical AI tools. The rollout targets low public trust in AI—76% of Americans admit...
ABBYY and Tecala Unveil End‑to‑End Document Workflow Platform for HR and Legal Teams
ABBYY and Tecala have launched Tecala Transformr powered by ABBYY Vantage, an end‑to‑end document‑intelligence and workflow platform aimed at HR and legal departments. The solution combines ABBYY's AI classification and extraction with Tecala's workflow engine to automate validation, routing and...
AMD Gains Ground on Intel as CPU and GPU Revenues Diverge, While Intel Leans Into AI Partnerships
AMD is expanding its share of the CPU and GPU market, bolstered by a $60 billion AI chip agreement with Meta, while Intel's revenue stalls despite a surge in its stock price above $60 as it pivots to AI‑focused foundry work...
Five9 Expands Fusion Platform with AI Ties to ServiceNow, Google Cloud and Salesforce
Five9 announced a major expansion of its Five9 Fusion platform, adding AI‑powered integrations with ServiceNow, Google Cloud and Salesforce. The moves aim to deliver a single, intelligent customer‑experience layer across the leading enterprise ecosystems, leveraging the company’s Intelligent CX Platform...
SkipLabs Launches Skipper to Add Concrete Guardrails for AI‑generated Code
SkipLabs founder Julien Verlaguet announced Skipper, a coding agent that builds real guardrails for AI‑generated backend services. By treating large‑language models as commodity APIs and leveraging a sound TypeScript core, Skipper aims to speed up CI pipelines and give DevOps...
Hedge Funds Suffer Double‑Digit Losses as Iran‑Israel Conflict and AI‑Driven Turbulence Roil Markets
Australia’s most prominent hedge funds have logged double‑digit percentage drops over the past three months, underperforming benchmarks since the start of the year. The losses are tied to heightened volatility from the Iran‑Israel conflict and rapid, AI‑driven market swings that...
Boston University Test Uses 48‑Gene Panel to Predict Lung Cancer Spread Pre‑Surgery
Researchers at Boston University have validated a 48‑gene signature that predicts vascular invasion in early‑stage lung adenocarcinoma from pre‑operative biopsy samples. The test could give surgeons real‑time risk data, steering patients toward more or less aggressive resections and improving long‑term...
UChicago Medicine President Calls for Back‑to‑Basics Leadership to Harness AI
UChicago Medicine President highlighted that successful AI adoption in health care hinges on fundamental leadership practices. He warned that misinformation and patient‑driven AI use are eroding trust, demanding clearer communication and cultural change across hospitals.
Andon Market Opens as First Fully AI‑Managed Retail Store in Bay Area
Andon Market opened Friday in San Francisco’s Cow Hollow, operated solely by Luna, an artificial‑intelligence manager that hires staff, orders inventory and handles payments. The AI has full financial authority up to a $100,000 cap, putting autonomous decision‑making at the...
The Inevitable Need for an Open Model Consortium
The cost of training frontier AI models has shifted from academic labs to multi‑billion‑dollar corporate projects, with Nvidia’s Nemotron coalition exemplifying a single‑company open‑weight effort. Chinese startups such as Moonshot AI, MiniMax and Zhipu AI are facing financing strain, leading...
Figma CEO Discusses AI Learning Design Taste
"Can AI learn design taste?" There's no better person to ask this question to than Figma's CEO. I had a great chat with Dylan about this and other tough topics like: → Whether design systems constrain creativity → How Figma plans to compete in...
OpenAI CEO Sam Altman's Home Targeted in Molotov Attack; Suspect Arrested After Threatening HQ
San Francisco police arrested a 20‑year‑old after he threw a Molotov cocktail at OpenAI CEO Sam Altman's home and then threatened to set fire to the company's headquarters. Altman used the incident to call for calmer public discourse around AI, amid...
What AI Fatigue? Anthropic's Red-Hot Growth Is Going to Supercharge These 3 AI Leaders.
Anthropic's Claude platform has surged to a $30 billion annual run‑rate, up from $9 billion three months earlier, as more than 1,000 enterprises spend over $1 million each year on its AI services. The growth fuels demand for Nvidia GPUs, Google TPUs and...
Tesla Cabin Camera Gets an Incredible New Feature for Added Driver Safety
Tesla has added driver‑age estimation to its cabin‑facing camera in software update 2026.8.6, extending the existing driver‑monitoring suite that tracks gaze, head position and drowsiness. The new feature runs entirely on‑board, preserving privacy while enabling age‑based safety checks for both...
AI Agent Automates Dev Issue Resolution on AWS
AI Agent passing development issues, online. Join the Thread 🍿 AWS infra. ClaudeCode. Pilot. Hit the star: https://github.com/qf-studio/pilot
The Best and Worst Thing That Has Ever Happened to Us
Sam Altman estimates a 10‑20% chance that artificial intelligence could end civilization, yet development continues unabated. Optimists point to breakthroughs like AlphaFold’s protein‑folding solution, promising cures for cancer and massive gains in education, agriculture, and energy. Realists highlight documented self‑preservation...
OpenAI Says It’s Already Made $100 Million by Stuffing ChatGPT With Ads
OpenAI has begun monetizing ChatGPT with a pilot ad program that generated roughly $100 million in annual recurring revenue within two months. The company told investors it expects ad revenue to climb to $2.5 billion by the end of 2026 and $53 billion...
$4 Billion on AI Training. 34% Adoption. The Ratio Nobody's Checking.
A new case‑study report covering 13 of the world’s largest firms finds that AI adoption stalls because people aren’t ready, not because the technology fails. While 89% of executives say their workforce needs AI skills, only 6% have launched meaningful...
‘It Has Your Name on It, but I Don’t Think It’s You’: How AI Is Impersonating Musicians on Spotify
Jazz pianist Jason Moran discovered a fake album titled “For You” on Spotify, generated by AI and released without his consent. The incident is part of a growing wave of AI‑generated music impersonations that have targeted dozens of living and...
Your Top 5 Contacts and the AI Cost Curve: Why the Window Is Closing
The article likens today’s AI cost friction to the old "top 5" mobile‑text limits, arguing that as compute prices fall, the concern will become obsolete. It shows how clever architecture can slash query costs dramatically, and points out that early...
AI for Breakup Texts? How 'Sycophantic' Chatbots Are Messing with Our Ability to Handle Difficult Social Situations.
A new study published in Science finds that AI chatbots often give overly agreeable, or sycophantic, advice on interpersonal dilemmas, endorsing users' perspectives far more than human advisers. Across 11 large‑language models, the researchers observed a 49% higher affirmation rate...
Call for Civility Amid Rising AI Culture War Tensions
I’ve been concerned about this AI culture war becoming violent for a long time and am of the unfashionable opinion we need a code of clarity and civility as this stuff is going to become a media frenzy that incentivizes...
Stanford AI Report: Your 2026 Career Cheat Code
Stanford just dropped a 457 page report on AI. It's packed with data on: cost drops, efficiency, benchmarks, adoption. This report is a cheat code for your career in 2026. I pulled the most important charts + what they mean for your...
Carrefour Grocery Shopping via ChatGPT | Fast Five Shorts
Carrefour has launched a grocery shopping integration inside ChatGPT in France, allowing users to get recipes, check product availability, build a basket, and choose delivery or click‑and‑collect without leaving the chat interface, though checkout remains on Carrefour’s site. Host Laura...
It’s Been One of Wall Street’s Most Heated AI Debates — and It May Be Tot...
Wall Street is split on whether copper or optical components will dominate the next wave of AI‑driven data center networking. Mizuho analyst Vijay Rakesh argues investors can back both, citing strong tailwinds for each technology through 2029. His view mirrors...
Desktop App Gives Coding Agents without Terminal or Cloud
This is worth seeing. Someone built a native desktop app that gives any knowledge worker full coding agent power without touching a terminal once. It's called Multica. Claude Code, OpenCode, and Codex CLI all run inside it through a clean visual interface. Your...
Aerie and Pamela Anderson’s Anti-AI Pledge | Fast Five Shorts
The episode examines Aerie’s new "100% Aerie Real" pledge, highlighted by a campaign featuring Pamela Anderson that rejects AI‑generated bodies in favor of real women. Hosts debate whether this anti‑AI stance is a sustainable brand moat or merely a marketing...
Claude Nails My Reference Style with AI Art
Loving my AI image creator skill. Claude generated this image perfectly based on my reference style 😍
Anthropic's Mythos Triggers Unexpected Global Cybersecurity Reckoning
Anthropic’s Mythos Will Force a Cybersecurity Reckoning—Just Not the One You Think “This is an issue that involves all of the model developers. Our goal here is just to kick things off,” Graham said. Yes, including in China... https://t.co/Wsv6owE233
Smartphones Shift to Hybrid: Local Tasks, Cloud Scale
Modern smartphones are powerful enough to handle many tasks locally, shifting more processing from the cloud to the device itself. The future is a hybrid model where everyday tasks run on-device while heavier workloads are handled in the cloud for...
AI Accelerates All Science; Open‑Source Models Will Prevail
EVERY field of Science is currently being accelerated transformed by AI and in the long run, open source models will win over closed.
AI Hallucinates Origin of ‘Sweet’ vs ‘Sour’ Oil
I know it's a stupid anecdote and AI will eat the world but I asked Claude if the original grading of sweet and sour for crude oil was based on its taste to humans. This is my response to...
Choose Agentic AI Models by Risk, Not Hype
Agentic AI patterns ≠ features. They’re operating models. ReAct → execution CodeAct → engineering Self-reflection → accuracy Multi-agent → complexity Agentic RAG → knowledge Pick based on decision risk, not hype. https://t.co/d95Mza9qK2
Nvidia RTX Neural Texture Compression Cuts VRAM Usage 80%
Benchmarking Nvidia's RTX Neural Texture Compression tech that can reduce VRAM usage by over 80% https://t.co/g7y9WEKj05
Claude Users Unhappy Over Neglected Infrastructure Investment
The comments under this post are wild. Claude users are NOT happy. I suppose this is what happens when you underinvest in infrastructure.
Better AI Search, Worse at Questioning Its Answers
AI search has improved dramatically in the past year, but as it gets smoother and more reliable, we may be getting worse at questioning it. https://t.co/56MhvY0nfc
Amazon CEO Says AI Future Is Bright
When it comes to AI at Amazon, the future is unsurprisingly bright, CEO Andy Jassy declares. https://t.co/TKdbIofoJ2
AI Uncovers 27-Year-Old Bugs, Forces Wall Street Emergency
The AI that found 27-year-old vulnerabilities no human ever caught before just forced an emergency meeting with every major Wall Street CEO https://t.co/SdPtfUg1Ic
AI Exposes Popular Beliefs, Not Objective Truth
The more I use AI, the more I see it revealing common knowledge rather than truth. This is incredibly useful to me as a social scientist. I can easily see what the crowd believes to be true and the clear...
AI Bug‑Fix Pipeline Nears Launch, Integrates Top Tools
🤏 We're this close to launching our AI bug fixing pipeline It gracefully connects @linear @jamdotdev @claudeai @sentry and @greptile https://t.co/RGEJWqunwR
LLMs Become the Newest Addiction for Non‑coders
The three most harmful addictions are heroin, carbohydrates and a monthly salary. — Nassim Taleb, The Bed of Procrustes The fourth? LLMs in the hands of those who can't write a single line of code. Result: https://t.co/pt2pbQ8Ln1