Today's AI Pulse
AI agents challenge cybersecurity frameworks, sparking budget hopes and governance concerns
AI agents are being deployed to test existing cybersecurity frameworks, prompting CISOs to see a path to unlock larger cyber budgets. At the same time, industry leaders warn that the lack of clear algorithmic authorization raises governance risks. The convergence of these trends highlights both the promise and the perils of ungoverned AI in security operations.
Also developing:
By the numbers: OneRobotics acquires Nanoleaf for $40M
Jack Dorsey Says His Employees Have Stopped Bringing Slide Decks to Meetings. Here’s What They Show Up With Instead.
Block CEO Jack Dorsey says employees have stopped using slide decks, now bringing AI‑generated prototypes to meetings. He argues prototypes provide greater realism and can be updated instantly, improving decision‑making. The shift follows Block’s AI‑driven restructuring that cut roughly 4,000 jobs, about 40% of its workforce, while gross profit continues to rise. Dorsey joins other tech leaders like Jeff Bezos and Elon Musk in moving away from traditional presentations.
Three YouTubers Accuse Apple of Illegal Scraping to Train Its AI Models
Three prominent YouTube creators—h3h3 Productions, MrShortGameGolf and Golfholics—have filed a class‑action lawsuit accusing Apple of illegally scraping their copyrighted videos to train its generative AI models. The complaint alleges Apple bypassed YouTube’s controlled streaming architecture, violating the Digital Millennium Copyright...
Secret Service Is Embedding AI Experts Across the Agency
The U.S. Secret Service is launching an internal AI program that embeds artificial‑intelligence specialists across its operations. Chief Information Officer Chris Kraft, a former DHS AI leader, says the small team will accelerate adoption of tools like license‑plate reading and...
Economists Expect AI to Lift Growth, but Not Transform It — at Least Not Yet
A new survey of economists, AI firms, forecasters and the public finds most experts assign a high probability to significant AI advances by 2030, yet they expect only modest near‑term macroeconomic effects, with a median unconditional GDP growth forecast of...
Ep 23: Cross-Stack Design and Tooling for Large-Scale Distributed AI Systems with Dr. Tushar Krishna, Georgia Tech
In this episode, Dr. Tushar Krishna discusses his evolution from network‑on‑chip research to designing large‑scale distributed AI systems, emphasizing the importance of cross‑stack co‑design that spans accelerators, memory hierarchies, and interconnect fabrics. He explains how predictable AI data‑flow patterns enable...
OpenAI's Safety Brain Drain Finally Gets an Explanation and It's Just Sam Altman's Vibes
OpenAI has dismantled its dedicated AI‑safety teams, prompting a wave of departures that helped spawn rival Anthropic. In a New Yorker profile, CEO Sam Altman attributes the exodus to a cultural mismatch, emphasizing rapid product development over traditional safety caution....
Hyperscaler Backlogs Show Growing Demand for AI Infrastructure
Data center capital expenditures surged 57% to $726 billion in 2025 and are projected to grow over 50% in 2026, pushing total spend past $1 trillion. The four hyperscalers—Amazon, Google, Meta and Microsoft—raised capex 76% year‑over‑year, with Amazon spending $131 billion in 2025...
The 15 Hottest AI Data And Analytics Companies: The 2026 CRN AI 100
CRN’s 2026 AI 100 spotlights 15 data‑management firms powering the surge of AI agents and generative models. Databricks announced a $1.4 billion annual revenue run rate for its AI suite, while Alteryx, ThoughtSpot, and others unveiled new agentic platforms that embed industry‑specific...
Can Artificial Intelligence Be Governed—Or Will It Govern Us?
The article draws a parallel between the post‑World War II nuclear arms control effort and today’s debate over artificial‑intelligence governance. It cites Marc Andreessen’s claim that AI regulation is akin to murder and highlights the backlash against Anthropic’s self‑imposed limits. By recounting...
Why AI Skills Are Becoming a Must-Have in Accounting Roles
AI proficiency has shifted from a nice-to-have to a baseline requirement for finance and accounting professionals. A Datarails analysis of 5,000 U.S. job postings shows AI or machine‑learning mentions rising to 30% for accounting roles, up from 18% a year...
NeuBird AI Closes $19.3M Round Led by Xora Innovation
NeuBird AI announced an oversubscribed $19.3 million funding round led by Xora Innovation, with participation from Mayfield, StepStone Group, Prosperity7 Ventures and Microsoft’s M12 fund. The capital will accelerate product development, global go‑to‑market expansion and broaden access for DevOps, SRE and...
Startup Approved to Let AI System Prescribe Psychiatric Medication
Legion Health, a San Francisco startup, received Utah regulatory approval to let its AI chatbot renew psychiatric prescriptions for a limited set of antidepressants such as Prozac and Zoloft. The system can only prescribe drugs previously authorized by a human...
IBM, Arm Target Enterprise AI With Mixed-Architecture Approach
IBM and Arm announced a partnership to run Arm‑native applications on IBM Z mainframes and LinuxOne servers through a shared software layer and virtualization. The solution lets enterprises deploy AI workloads across both architectures without rewriting code, preserving the uptime,...
We Start Tomorrow at 9am
The author launches a 14‑day, AI‑enhanced content program that promises to build a personal‑brand ecosystem with just 1‑2 hours of daily effort. Participants receive daily modules, starting tomorrow at 9 am MST, plus a pre‑launch "day 0" module designed to jump‑start the...
Gemma 31B Runs Locally: High Quality, Very Slow
got Gemma 4 31B on my mac... at ~10k tokens input it does struggle quite a bit to give me a response, the fans are getting pretty loud. Finished answering in 6 min 52 sec. Quality wise, it's good. Super slow, but...
Microsoft Says Copilot Isn't Just 'for Entertainment Purposes' After Its Terms of Service Language Goes Viral
Microsoft announced it will revise the Copilot Terms of Use after the "for entertainment purposes only" clause went viral on social media. The wording, a relic from Copilot’s early days as a Bing search companion, no longer reflects how the...
Tailor Your Prompt Style to Each AI Tool
Prompting isn’t one-size-fits-all. 🧠 This cheat sheet nails it: ChatGPT = instructor style ✅ Perplexity = research analyst style 🔍 Grok = candid friend style 😄 Gemini = project planner style 📌 Match the prompt to the tool → better outputs, faster. 🚀 What’s your favorite prompt...
Flying at 34k Ft, Still Crunching AI Workloads
Things I’m doing while flying at 34,000 feet: * Fine-tuning on my DGX Station (SSH) * Running 8 concurrent @cursor_ai cloud agents * Replying to emails * Posting on X
The Skills That Matter Most in the Age of AI – with Aneesh Raman
In this episode, host and guest Aneesh Raman discuss how organizations must evolve like massive startups to thrive in the AI era, granting employees autonomy to experiment with AI tools. Raman emphasizes the importance of reading the room and bringing...
Perfect Prompt Isn't the Secret to Generative AI
Debunking the biggest myth in Generative AI: that the secret lies in the perfect prompt https://t.co/itmbaNzBwU via @DLAIgnite #SocialSelling #DigitalSelling #Sales #SalesTips #SalesLeader #Salesforce #Marketing #Leadership #MarketingStrategy #ArtificialIntelligence #TechNews https://t.co/hwwvwtodF0
YouTube Expands AI Likeness Detection, Backs NO FAKES Act
At @YouTube, we're expanding our automated likeness detection technology to help prominent figures manage unauthorized AI-generated replicas. We’re also proud to work alongside Congress in support of the NO FAKES Act to establish a federal property right that protects your...
Rebellions AI Rings Up The Money To Rack Up AI Inference Systems
South Korean AI‑chip startup Rebellions AI secured a $400 million Series D round, bringing total funding past $850 million and valuing the company at $2.34 billion. The capital will fund its push beyond Korea, delivering complete inference systems—Rebel100 engines, RebelRack, and RebelPod—to global cloud,...
New Yorker Probes Sam Altman, Questions ChatGPT’s Sycophancy
The @NewYorker's new story about Sam Altman is very worth reading. @RonanFarrow & @andrewmarantz interviewed 100+ people and reviewed docs about various concerns with Sam & OpenAI. The incisive kicker might also make people ask if ChatGPT's sycophancy is a bug...
Fine‑tuning Gemma 4 at 34,000 Ft: Surreal Tech Adventure
Fine-tuning Gemma 4 on my DGX Station while flying 34,000 feet in the fair is a surreal feeling.
Netflix's Void AI Can Remove Objects From Video and Show How Scenes Evolve without Them
Netflix unveiled Void, a video object and interaction deletion framework that can erase objects from footage while realistically updating surrounding physics such as shadows, debris and water ripples. The system is trained on a synthetic paired dataset generated with Kubric...
AI, Values, and Trust: Rebuilding College Sports Media
Wrote a little bit this morning about AI in media and college sports administration...as well as values, service, and how to get fans (and readers) to trust you again. https://t.co/3aG47EVBNr
AI Lab Leaders Warn of Job Losses
The heads of the big AI labs continue to insist that their products are going to take all your jobs, and also pose various catastrophic risks
U. Florida Leads AI Education Task Force
The University of Florida has been appointed to organize and chair a new Florida K‑12 AI Education Task Force, funded by a $5 million donation from the Griffin Catalyst. The task force brings together 250 stakeholders from 39 school districts, charter...
AI Dev Tools Converge on Orchestration Across Environments
All the AI dev tools vendor are betting on orchestration. Where does that surface live? The terminal, like Claude Code? Sandalone desktop app like Codex? As part of the IDE like Antigravity and Cursor? @janakiramm takes a good look worth checking...
AI Endorses User Actions 49% More than Humans, Even Harmful
Stanford published a study in Science on LLMs that found that AI affirmed users' actions 49% more often than humans on average, including in cases involving deception, illegality, and other harms https://t.co/okEHzfkwTG #ArtificialIntelligence #Innovation #Technology #Tech #TechNews
‘Google AI Edge Eloquent’ Is an Offline, Subscription-Less Voice Dictation App
Google launched the AI Edge Eloquent app for iOS, a subscription‑free voice dictation tool that works entirely offline when enabled. The app transcribes speech in real time, cleans up the text after you stop speaking, and copies it to the...
Big Tech AI's Water and Energy Toll Sparks Scrutiny
Ignoring the environmental impacts of Big Tech’s AI push is lunacy. Investors & communities are right to question @amazon, @Microsoft, @Google, etc., on their surging water and energy usage Via @JessopReuters, @ValerieVolco, @Tech_Correspond, @Reuters https://t.co/eUtosmT1Ky
Meta to Open‑source some AI Models, Not All.
Scoop: Meta will still open-source AI models — just not all of them via @inafried https://t.co/fBELUAG0tj
Report Warns of AI’s Impact on Non-College Grads’ Careers
A Brookings Institution report warns that artificial intelligence could upend the career trajectories of the 70 million U.S. workers classified as "Skilled Through Alternative Routes" (STARs). Of these, 15.6 million occupy jobs highly exposed to AI, including 11 million "gateway" positions that traditionally...
Global Internet Needs Global Governance, Says EU Delegate
Too late to post about SXSW? One more highlight: I sat down with @RuthBajada (Deputy Head of the EU Delegation) and talked about AI, digital work, and building responsibly. Hot take: The internet is global. The rules shaping it should be too. Thank...
AI Will Serve as First-Line Support, Not Replace Therapists
Simply put, #AI won't replace therapists, but it will absolutely become the first line of support for many people who otherwise would have no access to care—so governance and guardrails matter more than ever. #health #healthcare https://t.co/aCaFxLV5kI
New Research Reveals Trade-Off Behind AI Vibe Coding Tools: Speed Vs. Understanding | Exclusive
A new research report from the University of Macau and ETH Zurich finds that AI‑driven vibe‑coding tools let children build games much faster, but they also short‑circuit deeper learning about game development. The study, which surveyed school‑age users of visual...
OpenAI News Overload? Essential Rundown for Newcomer Subscribers
too much openai news to keep track of. here's my rundown for paying Newcomer subscribers https://t.co/xBTfwBop9W
Voice AI Secures Over $1B Funding in January
Voice AI had a stellar first quarter, with startups getting more than $1.88 million in funding in voice and voice-related startups. January was the biggest month with over $1 billion in funding My story in the newsletter: https://t.co/xanR263SOd
AI-Written Code Can Beat Humans at Biomedical Analysis, some Studies Find. What Does that Mean for the Field?
A recent study published in Cell Reports Medicine shows that large language models can generate biomedical analysis code that matches or exceeds expert performance. Junior researchers, including a graduate student and a high‑school student, used simple prompts to produce accurate...
Less Work, Equal Pay: OpenAI Lays Out Its Vision for a World Reshaped by Superintelligence
OpenAI released a policy paper titled "Industrial Policy for the Intelligence Age" outlining early proposals for governments to manage the transition to superintelligence. The document suggests a public wealth fund that distributes AI‑driven returns to all citizens, a four‑day workweek...
The $100B Question: AI’s Appetite for Compute Is Rewriting the Rules of Tech
Recent financial filings from OpenAI and Anthropic reveal that AI development is becoming a capital‑intensive, compute‑driven industry. OpenAI projects $121 billion in compute spending by 2028, with inference expected to consume 65% of AI compute by 2026, while both firms report...
SageX AI Launches Unstructured Data Platform for Hedge Funds and Asset Managers – AI Data Transformation for Capital Markets
SageX AI has launched an unstructured data platform tailored for hedge funds and asset managers, promising to turn the 90% of unstructured data into AI‑ready, structured intelligence. The no‑code solution claims to cut data processing costs by up to 90%...
Onix Expands Google Cloud Collaboration Around Agentic AI and Data
Onix announced an expanded partnership with Google Cloud to accelerate enterprise‑scale AI deployments, committing to drive over $500 million in Google Cloud consumption and boost services revenue. The collaboration centers on Onix’s Wingspan platform, which blends data modernization with agentic AI...
The One Piece of Data that Could Actually Shed Light on Your Job and AI
AI’s potential to replace human labor has sparked panic, but economists warn that current tools—like task‑exposure scores from O*NET—are inadequate for forecasting job displacement. University of Chicago economist Alex Imas argues that without price‑elasticity data linking AI‑driven productivity gains to...
Sam Altman Watches Awkwardly As He’s Shown Bizarre ChatGPT Issue: “Uh, Maybe, Uhhh…”
OpenAI CEO Sam Altman was shown a TikTok clip where ChatGPT’s voice mode claimed a user’s two‑second run lasted over ten minutes, then defended the error. Altman brushed the glitch as a "known issue" and suggested it could take about...
CEA-Leti, CEA-List and PSMC Collaborate to Integrate RISC-V and MicroLED Silicon Photonics Into 3D Stacking and Interposer for...
CEA‑Leti, CEA‑List and Powerchip Semiconductor Manufacturing Corporation (PSMC) announced a partnership to embed CEA‑List’s RISC‑V processor IP and CEA‑Leti’s microLED silicon‑photonic chiplets into PSMC’s 3D‑stacking and interposer platform. The integration replaces traditional copper interconnects with short‑reach, high‑bandwidth optical links, delivering...
AI Safety Will Curb Crashes; Convenience Drives Driverless Adoption
Hopefully never. AI active safety features should hopefully make it nearly impossible for humans to crash while driving manually too. Convenience and saving on insurance, not a dystopian ban on driving, will be the incentive to move to a fully...
The Big Bang: A.I. Has Created a Code Overload
Financial services firms adopting AI‑driven coding tools such as Cursor have seen output explode from 25,000 to 250,000 lines of code per month, creating a backlog of roughly one million lines awaiting review. The surge has overwhelmed traditional security and...
ESET Previews New AI Security Features to Secure Chatbot Communications and AI Workflows
ESET unveiled a suite of AI‑focused security tools at RSAC 2026, slated for release later this year. The offering includes a browser‑level guard that inspects both prompts and LLM responses to block malicious links, scripts, and inadvertent data leaks. New...