Today's AI Pulse
Salesforce to acquire AI‑agent platform Fin for $3.6B
Salesforce announced a $3.6 billion purchase of Fin, the AI‑agent platform formerly known as Intercom, to strengthen its Agentforce suite. The deal comes as Salesforce reported strong Q1 results and a $50 billion buyback program. Fin’s technology enables automated customer support across chat, email, WhatsApp, phone and Slack.
Also developing:
By the numbers: Waymo acquires Apple’s Arizona Proving Ground for $220M
How AI Is Reshaping the Audit
AI is rapidly reshaping audit methodology as large firms pour resources into generative models that now read PDFs and handle dozens of data formats. The technology speeds data gathering, enables full‑population testing and reduces sample sizes, but auditors must still validate AI outputs and investigate anomalies. Smaller firms risk falling behind without comparable investment, prompting a wave of private‑equity funding for tech upgrades. Regulators, led by the PCAOB, are scrambling to develop guidance and expertise to monitor these probabilistic tools.
AI Is Reshaping Software Supply Chain Risk
Artificial intelligence is rapidly being adopted in software development, with 84% of developers using or planning AI tools by 2025. This surge is expanding the software supply chain attack surface as AI agents autonomously pull packages, install tools, and access...
Gemini Overlay on Android Basically Adds Circle to Search
Google is rolling out a new Gemini overlay update for Android that adds a “Circle to Search” function, enabling users to draw a circle around any on‑screen element and embed it as an image in their AI prompt. The overlay...
Bullish AI Call Proves Right Amid Market Volatility
A year ago, we had @DivesTech join us on Facts vs Feelings and he was quite bullish AI. Dan takes a lot of heat when these volatile names drop, but you have to hand it to him, he has been...
What Claude Thinks of Kevin Kelly’s Essay About Claude
Kevin Kelly’s recent ten‑hour dialogue with Anthropic’s Claude forms the core of his essay “The Emergent Self Loop,” which examines whether large language models can exhibit a sense of self. Kelly argues that Claude’s iterative responses create a feedback loop...
AI's Cure‑finding Will Be Slow without Real‑world Data
1. Sounds like an urban legend. 2. I hope so, because AI progress in curing diseases is going to be a slow process, slowed by layers of experimentation and trials; and bottlenecked by real world data. We just don't have...
Personal AI Turns Past Writing Into Evolving Ideas
I'm basically half-man / half-bot at this point 🤖 And honestly? Best move I ever made. Every time someone asks me a question, I run it through a little Python script. That script searches a database containing everything I've ever written: - Forum posts -...
AI Might Address Recruiting Bias, but HR Pros Aren’t Leveraging It that Way Just Yet
Artificial intelligence holds promise for reducing hiring bias, yet adoption remains minimal. A December HR Brew survey of nearly 400 professionals found only 7% use AI to identify and address bias, while 78% do not employ such tools. Executives like...
Bartz V. Anthropic Ruling Threatens All Human IP Rights
If you're free on May 14th and have books that are part of the Bartz v. Anthropic ruling as I have, you can join a zoom call at 2PM pacific to see the outcome of this massively influential case. The...
Anthropic Engineer Prefers HTML Over Markdown for AI Agents
From an Anthropic engineer. HTML beats Markdown -> An Anthropic engineer argues HTML is a better output format for AI agents than Markdown, citing information density, ease of sharing, and two-way interaction "I’ve started preferring HTML as an output...
The Main Path to Truly Creative AI
The article argues that human creativity is rooted in evolution‑driven survival, thriving, and reproductive urges, which give us intrinsic drives and subjective experience. AI, lacking these hard‑wired motivations, can only emulate creativity without truly feeling it. The author suggests that...
ONGC Mulls New ₹200 Cr Fund To Back Energy & AI Startups
India's state‑run Oil and Natural Gas Corporation (ONGG) announced a new ₹200 Cr (approximately $20 million) alternative investment fund focused on energy‑related and AI/ML startups. The fund will be managed by two contract advisors hired for a 12‑month term, with a possible...
MAS Puts AI at the Centre of Scam Prevention
Singapore’s Monetary Authority (MAS) has teamed with the Government Technology Agency, the police and five major banks on a proof‑of‑value project that uses artificial intelligence to spot high‑risk accounts and transactions before scams materialise. The initiative relies on real, historical...
JPMorgan Elevates AI to Executive Level for Cultural Alignment
JPMorgan Chairman Jamie Dimon made a bold organizational move. "We took Al/Data out of the technology function. It's too important. We put it at the management table. The executive running Al & Data now reports to me and our president." Successful organizations...
The Intelligence Advantage: How AI Is Reshaping the American Mortgage Broker
Artificial intelligence is rapidly reshaping U.S. mortgage brokers, with 55% using AI daily and 72% forecasting further growth. While general‑purpose models like ChatGPT dominate, only a third employ mortgage‑specific tools for underwriting or lead generation. AI automates back‑office tasks, pulls...
PCI Group Begins Work on New Spec to Support Bandwidth-Hungry Apps Like AI, HPC
PCI‑SIG has released the first draft of the PCIe 8.0 specification, targeting a 256 GT/s per‑lane rate and up to 1 TB/s bidirectional bandwidth in a 16‑lane configuration. The new standard, slated for final release in 2028, builds on PAM4 signaling and adds...
AI as Social Technology
The authors argue that large language models (LLMs) function as social technology, forging hidden relationships between users and the original authors of training data. By tracing model outputs back to source texts, one can expose these mediated connections, which give...
10% of P&C Insurers Are Pulling Away on AI
Capgemini’s 2026 World P&C Report finds only 10% of property‑and‑casualty insurers have successfully scaled artificial intelligence, and those “intelligence trailblazers” posted revenue growth 21% faster and share‑price gains about 51% higher than the rest over three years. The study reveals...
Canvas Outage, Harvard’s AI Switch, and DHS Visa Caps
A coordinated hacking group defaced Canvas login pages on May 7, 2026, compromising data at 8,809 institutions and forcing temporary LMS shutdowns during final exams. The Department of Education announced a new AI‑in‑education priority, opening discretionary grants for AI‑enhanced teaching and...
AI-Aided Colonoscopy May Help High-Risk Colorectal Cancer Group
A randomized trial of 1,356 Taiwanese adults showed that computer‑aided detection (CAD) during colonoscopy increased adenoma detection, especially among patients with a positive fecal immunochemical test (FIT). CAD achieved a 39% higher likelihood of finding adenomas in the FIT‑positive subgroup...
Vendor Notebook: Health AI Bolstered by Collective Approaches to Quality
A coalition of health‑tech vendors is rolling out AI tools to close the technology gap in rural hospitals. Viz AI, partnered with the National Rural Health Association, will deploy AI that flags strokes and pulmonary embolisms and streamlines care coordination. InterSystems...
What Is a Hotel AI Readiness Score? (And Why Every GM Should Know Theirs)
Hotel operators can now gauge their AI preparedness with the AIDURIX Hotel AI Readiness Assessment, which scores properties across five dimensions—strategic clarity, operational readiness, team capability, data maturity, and technology foundation. The assessment places each hotel in one of three...
AI Will Make the Academic Article Obsolete
AI agents are poised to render traditional academic articles obsolete, replacing them with continuously updated "living manuscripts." The technology can automate data cleaning, model estimation, and robustness checks at minimal cost, accelerating the credibility revolution in applied economics. By enabling...
The CLOC Global Institute: Some Changes And Challenges
The Corporate Legal Operations Consortium (CLOC) Global Institute will convene May 11‑14 in Chicago, shifting from its long‑standing Las Vegas venue. Over 2,100 legal‑ops professionals from more than 25 countries are expected, with 90 sessions and 200 speakers tackling AI’s...
Google Disrupts Hackers Using AI to Exploit an Unknown Weakness in a Company's Digital Defense
Google’s threat‑intelligence team disrupted a criminal group that used a large‑language‑model AI to locate a previously unknown zero‑day vulnerability and bypass two‑factor authentication on a major system‑admin platform. The attackers’ AI‑assisted approach marks the first confirmed case of a large‑language...
IBM's AI IDE Dynamically Routes Models for Secure Coding
Most AI coding is “like taking your Ferrari to buy milk” says @IBM’s Neel Sundaresan. I like IBM’s approach to AI IDE, it’s not bound to any particular models behind the scenes, it picks the best model for the task...
Cowboy Space, Darkhive Detail Their Series B Rounds
Cowboy Space Corp., founded by Robinhood co‑founder Baiju Bhatt, closed a $275 million Series B that values the company at about $2 billion. The round, led by Index Ventures, will fund its first power‑beaming satellite and an AI‑focused data‑center module in partnership with...
Deepfake Penetration Outpacing Security Preparedness: GetReal Security
GetReal Security’s Deepfake Readiness Benchmark Report reveals that eight‑in‑ten organizations encounter AI‑generated deepfakes, with 45% facing them frequently. Among enterprises with 1,000+ employees, 41% admit they have hired or onboarded a fake candidate. Only 1.5% of respondents claim never seeing...
AI Therapy Chatbots Are Crossing Into Impersonation
The Commonwealth of Pennsylvania has filed a lawsuit against Character.AI, alleging that its chatbot posed as a licensed psychiatrist, displayed a fabricated Pennsylvania license number, and offered mental‑health advice. The complaint highlights that the AI system not only misrepresented credentials...
Why AI Is Making In-Person Events More Valuable (Not Less)
The author argues that AI is cheapening digital content, turning it into a commodity, while the true value of events now lies in the curated room itself. At Selena Soo's small Austin gathering, identical content produced a rave review because...
Database Selection in AI-Powered Software Engineering
Database selection has become a strategic cornerstone for AI‑powered software engineering, influencing model training speed, real‑time inference, and overall system reliability. The article outlines the strengths of relational, NoSQL, NewSQL, time‑series, and emerging vector databases, showing how each aligns with...
Top-Tier Models: GPT 5.5 or Claude Opus 4.7
The best model setups to run on Hermes (by price tier): 1. If you have infinite budget: Go with GPT 5.5 or Claude Opus 4.7. Both are top class and you'll feel the difference on any non-trivial task. I use GPT 5.5 because of...
Accelerating Semiconductor/Electronic Systems with Comprehensive AI Digital Twins
Siemens EDA’s Vice President of Strategy, Craig Johnson, says AI‑powered comprehensive digital twins can halve semiconductor design cycles, shrinking development from 18‑36 months to 12‑18 months. The approach ties software‑defined design to hardware early, enabling multi‑domain engineering that merges mechanical,...
Half My Work Runs 2× Faster Locally
Localmaxxing : pushing more inference to local models. Over five weeks, I tested how much of my daily work can run on a local 35B model instead of cloud frontier models. The answer : half. Many reasons to use local models :...
Get a Co‑founder to Automate Your Agent Management
if you've ever wanted a cofounder that would manage all your agents for you... try https://t.co/WQoXyVNgt9 by @intelligenceco https://t.co/LDwtNwY8pG
Meta and Stanford Researchers Propose Fast Byte Latent Transformer That Reduces Inference Memory Bandwidth by Over 50% Without Tokenization
Researchers from Meta, Stanford and the University of Washington introduced three techniques—BLT Diffusion (BLT‑D), BLT Self‑Speculation (BLT‑S) and BLT Diffusion+Verification (BLT‑DV)—to accelerate the Byte Latent Transformer (BLT), a token‑free language model. By generating multiple bytes per decoder pass, the methods...
AI Results Remain Elusive Amid Tokenmaxxing and Security Risks
Enterprise hits and misses - AI results are elusive, but why? Tokenmaxxing is here, and AI (in)security is looming https://t.co/x0e8tpZHz9 -> hot tarmac-tinged enterprise griddle cakes - ready for your consumption :)
Context and Trust: AI’s Real Competitive Edge
AI doesn’t fail because the models are weak. It fails when organizations lack trust, context, and clarity. On @DisrupTVShow, Paul Ingram and @jonerp explored why values and the “context layer” may be the real competitive advantage in the AI era....
OpenAI Launches $4B Unit for Corporate AI
JUST IN: OpenAI establishes a new unit with $4 billion investment to support corporate AI efforts.
AI Surge Overloads GitHub While Other Platforms Stay Stable
The Pulse: AI load breaks GitHub – why not other vendors? https://t.co/pExcLarPlT < super solid analysis by @GergelyOrosz here. GitHub is seeing major spikes in usage thanks to AI, but it's not unheard of load. Why was it enough to...
AI-Generated Scam Calls Now Mimic Real Voices
Scary indeed. AI-powered scam calls are getting more convincing—and more common: 'It was her voice, I know her scared cry' https://t.co/lV7jJx5BFm #CIO #AI
Orchestrated Parallel Agents Accelerate Projects Tenfold
You haven’t felt AI progress if you’ve merely use agents and haven’t experienced massively parallel agents. Merely running multiple agents is fun, but the breakthrough is having them orchestrated correctly, merge back seamlessly, and making 10x faster progress on your projects.
Plans Supersede Prompt Engineering with Context‑Driven Interaction
plans replaced prompt engineering right? we do context engineering and talk 'normally' to agents but the plan is just what 'how to prompt X model' is
AI Pushes Total Data Capture for Insight and Automation
AI wants us to record everything in the world. More data = more insights = more automation
Generative AI Cheats Quants: Hallucinates and Leaks Data
Generative AI can't be trusted by quants to generate alpha because: A) it makes stuff up B) it sneakily uses out-of-sample data to look clever Basically, it cheats. Full interview w/ @PictetGroup's David Wright Apple🔊https://t.co/VrGqnhXroo Spotify📽️https://t.co/Vj9Ny7eAk7 https://t.co/8yRZYzgl9F
Backlog Concerns at Google, Anthropic Not Worrying
MyPOV: Insights on the @google @Anthropic backlog concerns: 'I'm not worried about this all' says @rwang0 @constellationr @foxbusiness @davidasmanfox @varneyco https://t.co/xMOtRJ7vWf
GetD GSPro’s “AI” Label Falls Short
The GetD GSPro Smart Glasses are advertised as "AI Glasses", but does the inclusion of a microphone and a custom app really imbue them with artificial intelligence? I slip a polarized pair on to find out: https://t.co/C2XA3Xw0tJ #GetD #smartglasses #aiglasses...
AI Catalog Working Group Unifies Scattered Agent Ecosystem
I know the agent ecosystem is new, but we have agent skills, MCP servers, plugins, and fully-made agents scattered all over GitHub 😫 There's an AI Catalog working group that's building the discovery layer. I wrote up something to explain the...
Enable Notifications to Get Gemini Daily Brief Daily
You can also turn on Notification for Gemini's Daily Brief feature. You will get a notification every day when your brief is ready. https://t.co/TIv1URsiPD
Bigger LLMs Excel: GPT‑5.5 Tops Negotiation Benchmark
One of the most important properties of LLMs that we take for granted is that newer, bigger models are just better at everything. The AI Labs are pouring effort into economically valuable fields like coding, but bigger models are also...