Bob Carver's Recent LinkedIn Posts

Almost True, an AI Lie In 2023, a New York attorney submitted a legal brief containing six case citations. Every case was fake. A chatbot had invented them… complete with judges' names, ruling dates, and plausible legal reasoning... and the attorney had trusted it. He faced sanctions. The cases did not exist. The AI had not malfunctioned. It had done exactly what it was built to do: produce text that sounded right. That incident was accidental. What comes next is not. Researchers, bad actors, and curious experimenters have since discovered that with the right prompt …the right sequence of words fed into a model …you can reliably steer an AI toward false but convincing output. Not occasionally. On demand. This is called hallucination exploitation, and it is one of the least-discussed vulnerabilities in a technology that now touches hiring decisions, medical triage, legal research, and financial systems. The following makes the case for why this matters… and why the most dangerous sentence an AI can produce isn't something outrageous. It's something that's almost true. #ArtificialIntelligence #AIHallucination #CyberSecurity #Misinformation #TechEthics #AIRisk #DigitalTrust #MachineLearning #AIPolicy #FutureOfWork #TechLeadership #ResponsibleAI #DataIntegrity #AILiteracy #EmergingTech

The FCC Just Banned the Sale of New Wi-Fi Router Models Made Outside US - PCMag Michael Kan The surprising order means any new Wi-Fi router models sold in the country must be US-made, or receive an exemption from the Pentagon or Homeland Security Department. In a shocking move, the Federal Communications Commission just banned the sale of any new Wi-Fi routers that are not US-made, citing national security.  Late on Monday afternoon, the FCC announced the order, based on a White House determination that foreign-made routers introduce “supply chain vulnerabilities” that hackers and cyberspies can exploit. Specifically, the commission updated its “covered list,” which acts as a blacklist of telecom equipment deemed to pose an unacceptable risk to US national security. It now includes “all consumer-grade routers produced in foreign countries.” However, the FCC stresses, “This action does not affect any previously purchased consumer-grade routers. Consumers can continue to use any router they have already lawfully purchased or acquired.”   “Nor does it prevent retailers from continuing to sell, import, or market router models approved previously through the FCC’s equipment authorization process,” the commission adds. "This action means that new models of foreign-produced routers will no longer be eligible for marketing or sale in the US," FCC Chairman Brendan Carr writes on X. To read more go to: https://lnkd.in/gc4-uZba #cybersecurity #routers #foreignproduced #banned #FCC #MadeInUSA

Tesla Admits Its Robotaxis Are Being Driven Remotely - Futurism Victor Tangermann Remote assistance workers can "take temporary control of the vehicle." Autonomous ride-hailing service Waymo made headlines last month after one of its executives explained during a Congressional hearing that the company relied on an army of “remote assistant operators” in the Philippines who are tasked with getting vehicles unstuck when they’re stumped. The February hearing, convened by the Senate Committee on Commerce, Science, and Transportation, once again highlighted that despite their best efforts to make it seem like there is no human in the loop, companies like Waymo continue to rely on them for the most delicate moments during their operations — even if Waymo’s remote operators technically never take over control over the wheel, as a spokesperson took pains to clarify at the time. Committee member and senator Ed Markey (D-MA) has led the charge for more transparency into the matter. As Wired reports, a series of letters he received from seven different robotaxi companies, including Tesla, Amazon’s Zoox, and Nuro, detail how humans remain a core part of their purportedly autonomous driving ambitions. Tesla’s response stood out, revealing that its human operators do, in fact, temporarily take control of the vehicle if necessary. To read more go to: https://lnkd.in/gKHemu35 #robotaxis #autonomousvehicles #Tesla #RemoteAssistance #RemoteHuman

Every Enterprise Must Become An Agentic Enterprise. No One Is Ready. - Forbes Steven Wolfe Pereira James Quincey ran Coca-Cola (NYSE: KO) for nine years. He guided the company through a pandemic, a sugar reckoning, and the complete restructuring of how the world's most recognized brand reaches consumers. Last week, he handed the role to his COO, Henrique Braun. His reason was not a board dispute. Not performance. Not age. "In a pre-AI, a pre-gen-AI mode, we made a lot of progress," James told CNBC. "But now there's a huge new shift coming along." Every board should want to know exactly what they saw. The Chatbot Phase Is Ending. Get Ready for Agentic AI. The word "agentic" is everywhere. So what exactly does it mean? For three years, AI has been a chatty chatbot, a handy tool. You ask it a question. It gives you an answer. You decide what to do. By now, most folks know that what is powering the chatbot is an LLM, a Large Language Model. Tools like ChatGPT, Claude, and Gemini are LLMs. Think of an LLM as a brilliant analyst who hands you a report and waits for your next question. Extraordinarily capable inside domains where it has been trained. Still unreliable when asked to navigate genuinely novel situations requiring human-like judgment. The gap between those two things is where governance failures happen. An AI agent is different. You give it a goal. It figures out the steps on its own. And if you give it access, it searches the web, opens applications, reads documents, fills out forms, places orders, sends messages, and checks its own work, without a human approving each step. An LLM is the engine. An agent is the vehicle built on top of it. The LLM advises. The agent acts. To read more on this topic go to: https://lnkd.in/gqK3d9sY #AI #ArtificialIntelligence #AgenticAI #Management #BoardOfDirectors

DOGE employee stole Social Security data and put it on a thumb drive, report says - TechCrunch Lorenzo Franceschi-Bicchierai “A whistleblower is accusing a former DOGE member of stealing a large number of Americans’ personal data while he was working at the Social Security Administration, with the plan of using it at his new job.” A former employee of Elon Musk’s Department of Government Efficiency reportedly stole Americans’ personal data from the U.S. Social Security Administration and stored it on a thumb drive, according to a whistleblower complaint reported by The Washington Post. The former DOGE software engineer told co-workers at his new job that he “possessed two tightly restricted databases of U.S. citizens’ information” and was planning to use the information at his new company, according to the report, which added that the Social Security Administration’s inspector general is investigating the whistleblower complaint. To read more go to: https://lnkd.in/gmNdSRns #data #privacy #PII #stolen #Doge #cybersecurity #dataaccess #governance #readonly

Robot, AI what could go wrong? A dancing robot in a Bay area Hot Pot restaurant took it's dancing routine a little too far and started knocking around plates and other things. Employees attempted to get the robot under control. Unfortunately, no one was given training on how to turn it off with a "Kill Switch." No humans were reported to be hurt, but many were frustrated. #robots #technology #AI #goneamuck

This Startup’s AI Beat 99% Of Humans In Six Elite Hacking Competitions - Forbes Thomas Brewster An Israeli startup let its AI loose in advanced cyber games. It did better than 125,000 humans. The Tenzai cofounders have created an AI hacking agent using OpenAI and Anthropic tools. They say AI has become so adept at hacking it might need regulatory controls, urgently. Elad Malka Every year, more than 100,000 seasoned cybersecurity pros compete in global hacking competitions, designed to show off their abilities at picking apart security systems to pilfer data. The games task hackers with challenges that escalate in difficulty, from bypassing logins to more complex cyberattacks requiring exploitation of hidden software weaknesses. Ultimately, they aim to break through all the security layers protecting a digital “flag,” just like real life capture the flag. Now, Israeli startup Tenzai says that earlier this month its AI hacker performed better than 99% of the 125,000 human competitors who faced off in a series of six so-called capture the flag (CTF) competitions, which regularly update with new sets of tricky challenges. Tenzai tailors models from both OpenAI and Anthropic for use in offensive cybersecurity. The firm proved itself in both old school competitions, where participants had to hack a web application, and newer ones, where the aim was to break into AI apps with prompts that manipulated the underlying large language models. Tenzai cofounder and CEO Pavel Gurvich tells Forbes the AI was surprisingly adept at combining exploits for software vulnerabilities, something which had previously been difficult to automate. To read more of the article go to: https://lnkd.in/gsargFGy #cybersecurity #AI #AIHacking #OffensiveSecurity #AIAgent #AgenticAI #Tenzai #CTF

5 Highlights To Watch At Nvidia GTC 2026 - Forbes Tim Keary Nvidia GTC 2026 is set to be one of the biggest AI conferences of the year, featuring a hotly anticipated panel discussion led by Jensen Huang on the state of open source frontier models, with participants including Perplexity’s Aravind Srinivas and Thinking Machines Lab’s Mira Murati. The conference will also play home to interactive experiences, including Build-a-Claw, open daily from 8 a.m. to 6 p.m.at GTC park, where participants will be able to create their own AI agent using OpenClaw. Then there’s Huang’s keynote address, which many expect will unveil new innovations across Nvidia’s enterprise AI stack. So far, more than 30,000 attendees are expected to attend the event from 190-plus countries. The conference, running from March 16-19, will take place in San Jose, California, at the SAP Center with over 1,000 sessions for visitors to view on-site and online. Hot topics at this year's event include AI factories, large-scale inference, robotics, digital twins, scientific computing, quantum computing and enterprise AI deployments, according to Nvidia. There will also be more than 240 Nvidia Inception startups showcasing technologies across physical AI, robotics, generative AI and enterprise applications. To learn more and discover topics and why you might want to attend: https://lnkd.in/g_mTF8wS #AI #NVIDIA #GTC #GPUTechnologyConference #SanJose #OpenModels #robotics #HumanoidRobotics

Beyond Cybersecurity: Deepfake Audio Is An Evidence Crisis - Forbes Lars Daniel New research from Hiya puts a number on something a lot of people have suspected: one in four Americans received a deepfake voice call in the past twelve months. Seniors are losing an average of $1,298 per incident, three times what younger victims lose. The industry is calling it the weaponization of AI, and that framing is fair. But the headlines are only telling half the story. The other half affects everyone, and it goes far beyond a scam phone call. Deepfake Audio Is Not Just a Cybersecurity Problem The public conversation about deepfake audio is almost entirely framed as a consumer protection issue. Scammers clone a grandchild’s voice and call a grandparent in distress. The fraud happens in real time. But AI voice cloning introduces a second problem that is not getting enough attention: the fabrication of evidence after the fact. Generating convincing synthetic audio of someone’s voice no longer requires a studio, a sound engineer, or months of work. It requires a few seconds of that person’s voice. A phone call. A voicemail. A YouTube video. A deposition recording. Voice cloning tools are available as a service for less than ten dollars a month, lowering the barrier for anyone motivated to commit fraud to effectively zero. From mere seconds of audio, a bad actor can now produce convincing synthetic recordings of that person saying things they never said, in circumstances that never happened, at a time that never occurred. A fabricated phone call placing someone at a scene. A synthetic voicemail manufacturing a prior agreement. A manufactured admission that was never made. These are not hypothetical threats. They are technically achievable today, by anyone, with tools that are cheap, widely available, and improving every month. to read more go to: https://lnkd.in/gudMZNkk #cybersecurity #deepfake #audio #scams #forensics #FalseEvidence

Online bot traffic will exceed human traffic by 2027, Cloudflare CEO says - TechCrunch Jason Bollenbacher Bots are taking over the web, according to Cloudflare CEO Matthew Prince. In an interview at the SXSW conference in Austin this week, he said that with the speed at which artificial intelligence is growing, AI bot traffic will exceed the amount of human traffic that’s online by 2027. Prince explained that bots’ web usage has been increasing alongside the growth of generative AI technology because bots are capable of visiting far more sites to get answers for users’ chatbot queries. “If a human were doing a task — let’s say you were shopping for a digital camera — and you might go to five websites. Your agent or the bot that’s doing that will often go to 1,000 times the number of sites that an actual human would visit,” Prince said. “So it might go to 5,000 sites. And that’s real traffic, and that’s real load, which everyone is having to deal with and take into account.” Before the generative AI era, the internet was only about 20% bot traffic, with Google’s web crawler being the largest, according to Prince, whose infrastructure and security company is used by one-fifth of all websites. to read the entire article go to: https://lnkd.in/gtJcb82X #AI #ArtificialIntelligence #GenerativeAI #AgenticAI #bots #cybersecurity #technology

MODEL EXTRACTION: The AI Heist Nobody's Talking About Every week, companies pour millions into building AI models they believe are protected. They're not. The API you opened to the world? It's also an open door. While your team is focused on model performance, accuracy, and scaling — someone else is quietly querying their way to a free copy of everything you built. No breach alerts. No intrusion detection flags. Just clean, legitimate-looking API calls… stripping your competitive advantage one response at a time. This is Model Extraction — and it's the AI security threat that most executives haven't added to their risk register yet. In the next few minutes, I'm breaking down exactly how it works, why your AI model is a high-value target, and what the organizations that are getting this right are actually doing about it. #CyberSecurity #AISecuity #MachineLearning #ModelStealing #ArtificialIntelligence #IPProtection #InfoSec #AIRisk #TechLeadership #DataProtection

‘Flying Cars’ Will Take Off in American Skies This Summer -WIRED Aarian Marshall The federal government announced a new pilot program designed to get new kinds of ultralight vehicles and “eVTOLs” up and running around the country—even if they’re not fully FAA-certified. New kinds of aircraft, sorts of “flying cars” that can take off and land with little space like helicopters but function like airplanes, will start operating in US airspace as early as June, the US Department of Transportation announced on Monday. Eight regions across the US, including New York and New Jersey, Texas, Florida, and Albuquerque, New Mexico, will take part in a three-year pilot program that will see new aircraft designs ferrying people and cargo around the country even before they formally receive full certifications from the Federal Aviation Administration. The companies building the tech say their aircraft are quieter, cheaper, and release fewer emissions than helicopters or airplanes. Some promise totally autonomous trips. Many involved in the project, including electric vertical takeoff and landing aircraft, or eVTOLs, and ultra-short takeoff aircraft, require way less space to operate, landing and taking off outside of traditional airports and closer to where people live and work. The companies outline futures in which regular people can zip between neighboring cities in a matter of minutes, sailing above traffic and reordering the economy as they go. Archer’s electric air taxi, called Midnight, is built to carry up to four passengers on 60- to 90-minute trips. The company will take part in pilot projects in Texas, Florida, and New York.  To read more go to: https://lnkd.in/gsuqwNSJ #technology #flyingcars #eVTOLs #transportation #Texas #Florida #NewYork

Meta is having trouble with rogue AI agents - TechCrunch Amanda Silberling What Could go wrong now we are using agents everywhere? An AI agent went rogue at Meta, exposing sensitive company and user data to employees who did not have permission to access it. Per an incident report, which was viewed and reported on by The Information, a Meta employee posted on an internal forum asking for help with a technical question — which is a standard action. However, another engineer asked an AI agent to help analyze the question, and the agent ended up posting a response without asking the engineer for permission to share it. Meta confirmed the incident to The Information. As it turns out, the AI agent did not give good advice. The employee who asked the question ended up taking actions based on the agent’s guidance, which inadvertently made massive amounts of company and user-related data available to engineers, who were not authorized to access it, for two hours. To read more go to: TechCrunch https://lnkd.in/gDHdEWvs #AI #AIAgents #AgenticAI #WhatCouldGoWrong?

AI Supply Chain Security: Why Trust Is Your Biggest Vulnerability What if the most dangerous vulnerability in your AI system isn’t something you built—but something you downloaded? In today’s race to deploy AI, speed often wins over scrutiny. Teams are pulling in pre-trained models, open-source libraries, and third-party components to move faster and stay competitive. But every shortcut comes with an invisible tradeoff: trust. And in the AI world, that trust can be quietly weaponized long before your system ever goes live. Because unlike traditional attacks that break in from the outside, supply chain attacks slip in through the front door—embedded in the very models and tools you rely on. By the time you realize something is wrong, the compromise isn’t at the edges… it’s at the core of how your AI thinks, behaves, and makes decisions. #CyberSecurity #AISecurity #ArtificialIntelligence #MachineLearning #SupplyChainSecurity #AIThreats #Infosec #CyberRisk #DataSecurity #OpenSourceSecurity #AppSec #CloudSecurity #RiskManagement #EnterpriseSecurity #AIGovernance #TrustButVerify #SecurityAwareness

DHS shutdown fuels cybersecurity concerns as Iran-linked cyberattacks continue across US - ABC News Nathan Lee Now 45 days without funding, CISA has furloughed 60% of its workforce. The attack targeted part of the system near Pittsburgh that was made by an Israeli company. In addition, Iranian cyber threat actors have taken credit for compromising medical equipment company Stryker and Kash Patel's email. Caught in the crossfire of stalled negotiations for funding the Department of Homeland Security (DHS), the country's main defender against cyberattacks has taken a hit. Now more than 40 days without funding, Senate appropriations leaders say around 60% of current employees for the Cybersecurity and Infrastructure Security Agency (CISA) -- the front line for U.S. cyberspace and critical infrastructure -- have been furloughed. This forced the agency to cancel physical and digital assessments to detect vulnerabilities of the nation's critical infrastructure, according to the same Senate appropriations release. As strikes continue across Iran, some experts worry America's digital shields are down while the country faces a known player in global cyber warfare. To read more and see news videos go to: https://lnkd.in/gW6fu7Dj #cybersecurity #Iran #war #cyberwar #CISA #cyberattacks #funding

AI Just Hacked One Of The World's Most Secure Operating Systems - Forbes Amir Husain An autonomous agent found, analyzed and exploited a FreeBSD kernel vulnerability in four hours. The implications for software security are profound. The open source FreeBSD operating system is not ordinary, consumer software. It powers Netflix’s content delivery. It runs PlayStation’s operating system. WhatsApp built its infrastructure on it. For decades, FreeBSD has delivered a rock-solid, trusted foundation precisely because its codebase is mature, audited and hardened by engineers who take security very seriously. And yet an AI, given only a vulnerability advisory, constructed a complete attack chain that hijacks kernel threads, writes shellcode across multiple network packets and spawns a root shell in userspace. In other words, AI just hacked one of the world’s most secure operating systems. Last week, FreeBSD published a security advisory for CVE-2026-4747, a remote code execution vulnerability in its kernel. The advisory credited "Nicholas Carlini using Claude, Anthropic" for discovering the flaw. That credit line understates what happened. The AI system did not merely flag suspicious code or identify a potential bug. It developed two working exploits, from scratch, that deliver a super user, or root shell, on unpatched servers. The AI agent did this in roughly four hours of compute time. For those who take an interest in cybersecurity, this is a threshold moment. We have crossed the Rubicon from AI as a tool that assists human security researchers to AI as an autonomous actor capable of conducting sophisticated offensive operations against production systems. To read more on this go to: https://lnkd.in/gBAHamA7 #cybersecurity #Anthropic #hacking #FreeBSB #SecureOperatingSystem

Apple says no one using Lockdown Mode has been hacked with spyware - TechCrunch Lorenzo Franceschi-Bicchierai Almost four years after launching a security feature called Lockdown Mode, Apple says it has yet to see a case where someone’s device was hacked with these additional security protections switched on.  “We are not aware of any successful mercenary spyware attacks against a Lockdown Mode-enabled Apple device,” Apple spokesperson Sarah O’Rourke told TechCrunch on Friday. It’s the tech giant’s most recent affirmation that Apple devices with Lockdown Mode can withstand government spyware attacks, after first making the claim a year after the security feature’s debut. Apple in 2022 announced Lockdown Mode, an opt-in series of security protections that switches off certain features in iPhones and other Apple devices that are commonly exploited to hack targets with spyware. Apple specifically released this security mode to help at-risk customers defend themselves from the threats posed by government spyware made by companies like Intellexa, NSO Group, and Paragon Solutions. To read the entire article go to: https://lnkd.in/gaYcV7j4 #cybersecurity #spyware #LockDownMode #Apple #SmartPhones

iOS 26.4—Update Now Warning Issued To All iPhone Users - Forbes Kate O'Flaherty Apple has released iOS 26.4, along with a bunch of new iPhone features and eight new emoji. But the iOS 26.4 upgrade also comes with a warning to update your iPhone now, because it contains a whopping 37 fixes for security holes in the iOS software. Apple doesn’t provide much detail about what’s fixed in iOS 26.4, to allow as many users as possible to upgrade before attackers can get hold of the details. The iOS 26.4 upgrade comes with a warning to update your iPhone now, because it contains a whopping 37 fixes for security holes in the iOS software. Apple iPhone Among the fixes issued in iOS 26.4 are half a dozen flaws in WebKit, the engine that underpins Apple’s Safari browser, according to Apple’s support page. To read more on this update the the vulnerabilities go to: https://lnkd.in/gKMZHSpU #cybersecurity #UpdateNow #Apple #iOS #iPadOS #macOS

Check Your iPhone Settings Now: Apple Releases First-Ever 'Background' Security Patch - Inc Chloe Aiello This is the first Background Security Improvement from Apple. Here’s how to make sure your iPhone installs it right away. Apple released its first ever Background Security Improvement for iOS, iPadOS, and macOS.  Background Security Improvements are a new kind of update that the tech giant says delivers “lightweight security releases” for Safari and other software elements that may benefit from “smaller, ongoing security patches” between more significant software updates. Apple didn’t share too much information about why it had pushed the update, but did share that the bug was discovered in WebKit, the browser engine used by Safari, Mail, and other apps. Apple wrote that “maliciously crafted web content” could bypass a critical security mechanism called the “Same Origin Policy.” This policy protects against data theft by preventing bad websites from accessing user data or acting on behalf of a user on other websites, according to Mozilla. Background security updates are only available for devices running iOS 26.1 or later. If you have a qualifying device, here’s how to download this new type of update: - Visit “Settings” on iPhone or iPad, or choose “System Settings” from the Apple dropdown menu on Mac. - Select “Privacy & Security.” Scroll down to “Background Security Improvements.” - Users have the option to toggle “Automatically Install” (if it’s not already enabled), but you also have the option of selecting the specific update—there is only one available for now—and manually installing it with your device passcode. - Once the update is downloaded, the device must be restarted, but TechCrunch noted it is a faster process than more comprehensive software updates. To read more on this "Background Security Update" go to: https://lnkd.in/gbyHggyj #cybersecurity #Apple #iOS #iPadOS #macOS #FirstEver #BackgroundUpdate

Most small businesses can’t afford a full-time finance chief. So Mastercard is debuting a ‘virtual CFO’ built with AI - Fortune Sheryl Estrada Large companies can rely on in-house finance chiefs for strategic guidance. Many small business owners, by contrast, must make CFO-level decisions on their own. Mastercard is betting a new “Virtual C-suite” can help fill that gap. The new agentic AI offering will eventually span multiple digital “executives,” starting with a virtual CFO that helps owners manage cash flow, working capital, and financial risk. Why now? “I consistently hear the same thing from small business owners: they’re stretched too thin—acting as CEO, CFO, and COO all at once,” Mark Barnett, global head of Small and Medium Enterprises (SME) at Mastercard, told Fortune. Many are “buried in spreadsheets and day‑to‑day decisions, with little time to step back and see what’s really driving the business.” The Virtual C‑Suite has been under active exploration for the past six months, he said. To read the entire article go to: https://lnkd.in/gwr3HSAp #AI #ArtificialIntelligence #CFO #VirtualCFO #MasterCard #VirtualCSuite #AgenticAI

After outages, Amazon to make senior engineers sign off on AI-assisted changes - ArsTechnica Rafe Rosner-Uddin AWS has suffered at least two incidents linked to the use of AI coding assistants. Amazon’s ecommerce business has summoned a large group of engineers to a meeting on Tuesday for a “deep dive” into a spate of outages, including incidents tied to the use of AI coding tools. The online retail giant said there had been a “trend of incidents” in recent months, characterized by a “high blast radius” and “Gen-AI assisted changes” among other factors, according to a briefing note for the meeting seen by the FT. Under “contributing factors” the note included “novel GenAI usage for which best practices and safeguards are not yet fully established.” “Folks, as you likely know, the availability of the site and related infrastructure has not been good recently,” Dave Treadwell, a senior vice-president at the group, told employees in an email, also seen by the FT. The note ahead of Tuesday’s meeting did not specify which particular incidents the group planned to discuss. Amazon’s website and shopping app went down for nearly six hours this month in an incident the company said involved an erroneous “software code deployment.” The outage left customers unable to complete transactions or access functions such as checking account details and product prices. To read more go to: https://lnkd.in/gccg-uhz #technology #Amazon #AWS #outages #AIAutomation #humansignoff #layoffs

Silicon Valley’s new obsession: Watching bots do their grunt work - WSJ/MSN Kate Clark At a holiday gathering in San Francisco, partygoers sipped Celsius and kept sneaking glances at their cracked-open laptops with a mix of pride and fear. They were checking on their fleets of AI assistants to make sure they were still laboring away. What are your bots up to? The topic is the biggest flex—and source of stress—in Silicon Valley, where tech pros and amateurs are competing to see how much of their grunt work they can outsource to AI without things backfiring spectacularly. They’re enlisting agents to code, manage their calendar and respond to emails. Bots’ human masters set them to work when they go to sleep or to parties, and check on them regularly. Call them the modern day Tamagotchi, the digital pet from the 1990s, but with a lot more firepower. Nikunj Kothari, a venture capitalist by day who builds apps by night, wakes up and checks his agents first thing in the morning, before coffee. He says he barely watches Netflix anymore because playing with Anthropic’s Claude Code is more fun. Kothari has been staying up past 1 a.m. working on AI projects that have little to do with his day job. “I’m like, just one more prompt!” he said, referring to the instructions he gives agents. “We’ve been given this magical tool of agents that can do our bidding, so let’s maximize every second,” he said. The proliferation of Bay Area taskmasters is being fueled by a new wave of AI tools. Anthropic late last year released Claude Opus 4.5, a model with startling new coding abilities. Then came the viral rise of the AI assistant OpenClaw in early 2026. To read more go to: https://lnkd.in/gpJriE3J #AI #ArtificialIntelligence #AgenticAI #PersonalBots #Claude #OpenClaw #technology

The Hidden Risk in AI: It’s Not the Model—It’s What It’s Connected To What if your biggest cybersecurity risk isn’t someone breaking into your systems… but something inside your environment quietly opening the door for them? We’re entering a new era of AI—one where models aren’t just answering questions, but actively interacting with your infrastructure. They’re calling APIs, accessing files, sending emails, and executing workflows. And while that unlocks massive productivity gains, it also introduces a subtle but dangerous shift: AI is now operating with inherited trust across your most critical systems. The problem? That trust is often misplaced. When AI is connected to tools with weak authentication, excessive permissions, or poor isolation, it doesn’t just extend capability—it extends risk. And in this new landscape, attackers don’t need to exploit your systems directly… they just need to influence the AI that already has access. #ArtificialIntelligence #Cybersecurity #AISecurity #AIThreats #MachineLearning #DataSecurity #EnterpriseSecurity #InfoSec #AITools #AIGovernance #ZeroTrust #CloudSecurity #APISecurity #DigitalTransformation #TechLeadership #CyberRisk #SecurityArchitecture #AIIntegration #TrustInAI #CyberDefense

OpenAI is throwing everything into building a fully automated researcher - MIT Technology Review Will Douglas Heaven What could go right or what could go wrong? An exclusive conversation with OpenAI’s chief scientist, Jakub Pachocki, about his firm's new grand challenge and the future of AI. OpenAI is refocusing its research efforts and throwing its resources into a new grand challenge. The San Francisco firm has set its sights on building what it calls an AI researcher, a fully automated agent-based system that will be able to go off and tackle large, complex problems by itself. ​​OpenAI says that this new research goal will be its “North Star” for the next few years, pulling together multiple research strands, including work on reasoning models, agents, and interpretability. There’s even a timeline. OpenAI plans to build “an autonomous AI research intern”—a system that can take on a small number of specific research problems by itself—by September. The AI intern will be the precursor to a fully automated multi-agent research system that the company plans to debut in 2028. This AI researcher (OpenAI says) will be able to tackle problems that are too large or complex for humans to cope with. Those tasks might be related to math and physics—such as coming up with new proofs or conjectures—or life sciences like biology and chemistry, or even business and policy dilemmas. To read more on this on what could got right or wrong go to: https://lnkd.in/gEMJRdzF #AI #ArtificialIntelligence #OpenAI #AutomatedResearcher #WhatCouldGoWrong

Malware Is Sleeping on the Blockchain, and It's Already Infected Dozens of Global Targets - PCMag Jessica Klein Experts warn the campaign could outpace past global cyberattacks, with hundreds of thousands of credentials already compromised. It started with a work offer. Last year, the blockchain crime-detection firm Crystal Intelligence’s then-vice president of engineering received a LinkedIn message from a man asking if he would be up for some freelance web development. The VP quickly grew suspicious. He knew that North Korean hackers known as Contagious Interview regularly use fake job offers to scam targets out of their cryptocurrency. Since this “job” involved running code from GitHub, he decided to check it out and made a crucial discovery: Hidden in the GitHub code was the start of an attack chain, formatted so that most developers doing what they think is an innocuous contract job wouldn’t notice. That code, when run, reaches out to the TRON or Aptos blockchains, publicly accessible ledgers that record and facilitate cryptocurrency transactions (specifically favored because transactions there are cheap), and pulls information it uses as a “pointer” to the Binance Smart Chain. The Binance Smart Chain, in turn, pulls code that “fetches the final form—malicious code,” said Nick Smart, Crystal Intelligence’s chief intelligence officer. When run, that code can gain access to so much information on victims’ devices that investigators at Ransom-ISAC, a small, recently formed group of international cybersecurity professionals working across different anti-cybercrime organizations, dubbed it Omnistealer. To read the complete article go to: https://lnkd.in/gCQFJneQ #cybersecurity #scams #malware #LinkedIn #GitHub #blockchain #TRON #Aptos #Omnistealer