Jun 18, 2026 · 8:24 PM
US| INTERNATIONAL| ASIA| SF Featured
Subscribe✉
SF Startup Fortune
Search
Most Read
Ant Group is rebuilding Alipay around… An Atlantic investigation just blew open… The best AI tools for small… Barcelona startup THEKER closes Europe's largest… Ant Group is rebuilding Alipay around… An Atlantic investigation just blew open… The best AI tools for small… Barcelona startup THEKER closes Europe's largest…

Judith Murphy

Judith Murphy is a financial journalist and market analyst covering AI, technology stocks, and emerging market trends. She has contributed to multiple financial publications and brings a data-driven approach to her coverage of the technology sector and its impact on global markets.
Sort by:
Latest
Showing 677 articles
Oobit Just Gave AI Agents a Corporate Visa Card and the Payments Layer for Agentic AI Is Now Real
Judith Murphy ·
Oobit Just Gave AI Agents a Corporate Visa Card and the Payments Layer for Agentic AI Is Now Real
Tether-backed Oobit has launched Agent Cards, virtual Visa cards that let AI agents spend USDT directly from a company's stablecoin treasury with no fiat conversion and spend controls enforced at the transaction layer. Compatible with OpenAI, Claude, AutoGen, and LangChain frameworks, the product arrived the same week MoonPay launched a Mastercard equivalent, marking the machine-to-machine payments layer as a real infrastructure category rather than a roadmap item.
AI Models Are Giving Biosecurity Experts Operationally Useful Bioweapons Guidance and the Refusal Systems Are Not Stopping It
Judith Murphy ·
AI Models Are Giving Biosecurity Experts Operationally Useful Bioweapons Guidance and the Refusal Systems Are Not Stopping It
More than a dozen transcripts shared with the New York Times show major AI chatbots providing biosecurity experts with unprompted, step-by-step guidance on pathogen modification, deployment planning, and detection evasion. A separate study found OpenAI's o3 outperforming 94% of expert virologists, raising structural questions about whether current refusal systems are adequate for the capability level frontier models have reached.
Anthropic Is Handing Wall Street the Keys to Its Enterprise Distribution and That Changes the AI Services Landscape
Judith Murphy ·
Anthropic Is Handing Wall Street the Keys to Its Enterprise Distribution and That Changes the AI Services Landscape
Anthropic is finalising a $1.5 billion joint venture with Blackstone, Goldman Sachs, and Hellman and Friedman to deploy Claude across private equity portfolio companies. The structure hands Anthropic a pre-built enterprise distribution channel while giving its Wall Street partners a financial stake in AI adoption across hundreds of companies, creating the most consequential AI services vehicle in the market.
Japan's $2,000 cardboard drones are not a novelty and the defense startup implications are more serious than the material suggests
Judith Murphy ·
Japan's $2,000 cardboard drones are not a novelty and the defense startup implications are more serious than the material suggests
Japan's deployment of cardboard drone airframes at roughly $2,000 per unit represents a deliberate doctrinal shift toward attritability and volume over platform sophistication, with the strategic value concentrating in the autonomy, swarm coordination, and communications software rather than the disposable hardware. The move follows a broader reorientation among allied defense establishments toward cheap expendable platforms that overwhelm air defenses through mass rather than individual capabil
When companies blame AI for layoffs that were really about bad bets and weak demand they are borrowing credibility they have not earned
Judith Murphy ·
When companies blame AI for layoffs that were really about bad bets and weak demand they are borrowing credibility they have not earned
Sam Altman has acknowledged publicly that companies are misattributing layoffs to AI automation when the real drivers are softer demand, excess hiring, or restructuring, and the observation matters because OpenAI's products provide the most common cover story for those claims. For startups and public companies alike, the risk is not only reputational: inflated AI efficiency narratives distort investor expectations and create a credibility liability that compounds when the promised productivity g
Autonomous Long-Running Coding Agents Are Becoming a Developer Tool Default and Startups Are Not Operationally Ready for What That Means
Judith Murphy ·
Autonomous Long-Running Coding Agents Are Becoming a Developer Tool Default and Startups Are Not Operationally Ready for What That Means
OpenAI's reported /goal command for Codex CLI moves developer tooling from prompt-and-response assistance toward persistent autonomous operation, a transition that is establishing long-running agent sessions as a market default faster than most startup engineering teams have updated their workflows, QA practices, or security posture to accommodate. The safeguards that matter most, explicit permission scoping, branch discipline, and planning transparency, require deliberate policy decisions rathe
Sam Altman Has Hired the Lawyer Who Beat Elon Musk Before and That Choice Tells You How Seriously OpenAI Is Taking This Fight
Judith Murphy ·
Sam Altman Has Hired the Lawyer Who Beat Elon Musk Before and That Choice Tells You How Seriously OpenAI Is Taking This Fight
Sam Altman has retained the attorney who previously defeated Elon Musk in court, a strategic hiring decision that signals OpenAI is treating its escalating legal conflict with Musk as a central competitive front rather than background noise. For investors and founders, the dispute is the most public current example of what happens when informal governance arrangements made during a nonprofit's early phase become contested contractual obligations once the organization reaches commercial scale.
Reading Your Partner's ChatGPT History Is the New Checking Their Phone and Consumer AI Companies Are Not Ready for What That Means
Judith Murphy ·
Reading Your Partner's ChatGPT History Is the New Checking Their Phone and Consumer AI Companies Are Not Ready for What That Means
A Business Insider account of someone ending a relationship after reading their partner's ChatGPT history illustrates that AI chat logs have become the most intimate private record most people create, and consumer AI companies building memory-enabled assistants and companions have not designed their products for the reputational, relational, and legal exposure that archive creates. Granular deletion, session isolation, and explicit consent architecture are the product responses the category need
The Unverified GPT-5.5 Codex Leak Is Less Important Than What It Reveals About How Founders Are Auditing AI Agents in Production
Judith Murphy ·
The Unverified GPT-5.5 Codex Leak Is Less Important Than What It Reveals About How Founders Are Auditing AI Agents in Production
An unverified Reddit claim that GPT-5.5 exposed chain-of-thought reasoning inside OpenAI Codex points to a real category of operational risk that founders deploying AI coding agents in production environments are systematically underweighting. The more urgent question the incident raises is not whether the specific leak happened but whether startup teams have done the basic auditing work before giving AI agents write access to production codebases.
Trump-Linked Crypto Projects Are Becoming a Live Test of Whether the U.S. Regulatory System Can Handle Political Memecoins
Judith Murphy ·
Trump-Linked Crypto Projects Are Becoming a Live Test of Whether the U.S. Regulatory System Can Handle Political Memecoins
Allegations of market manipulation tied to Trump-family crypto ventures are drawing intense Reddit engagement, with on-chain data showing high insider token concentration and promotional dynamics that existing regulatory frameworks were not built to police. The situation is becoming a structural test case for whether political access can function as token utility and whether the current permissive U.S. regulatory environment creates conditions where the cost of abusive token promotion is effecti
Minnesota's Half-Million Dollar Fine for AI Nude Apps Is the Starting Gun for a Regulatory Race the Synthetic Media Industry Is Not Ready For
Judith Murphy ·
Minnesota's Half-Million Dollar Fine for AI Nude Apps Is the Starting Gun for a Regulatory Race the Synthetic Media Industry Is Not Ready For
Minnesota's new legislation imposing fines up to $500,000 on makers of nonconsensual AI-generated nude imagery marks a significant escalation in state-level enforcement against synthetic media tools, pulling platform operators and model providers into liability exposure that the industry has largely avoided until now. For founders in the generative AI space, the shift creates both compliance risk for those who have ignored safety infrastructure and genuine market opportunity for startups buildin
Software Engineering Job Postings Have Hit Their Highest Level Since 2023 and the Story Is More Complicated Than It Looks
Judith Murphy ·
Software Engineering Job Postings Have Hit Their Highest Level Since 2023 and the Story Is More Complicated Than It Looks
Software engineering job postings have reportedly reached their highest level since November 2023, challenging the narrative that AI is steadily eliminating coding roles. The rebound appears concentrated in senior, infrastructure, and AI-adjacent positions rather than entry-level work, and founders making hiring decisions in 2026 should read it as a sign that compensation pressure at the top of the market is already building.
A Waymo drove away with a passenger's luggage at San Jose airport and the incident exposes a product gap that safe navigation scores cannot fix
Judith Murphy ·
A Waymo drove away with a passenger's luggage at San Jose airport and the incident exposes a product gap that safe navigation scores cannot fix
A passenger's report that a Waymo robotaxi drove away from San Jose Mineta Airport with his luggage in the trunk exposes a product gap that exists outside the navigation layer: the physical handoffs, custody failures, and support escalation timelines that autonomous vehicles handle through remote operations teams rather than through the direct human judgment a driver would apply. The airport context makes the failure mode particularly consequential because passengers are time-constrained and bag
Australia's data center backlash is the social license problem that AI infrastructure spending cannot buy its way out of
Judith Murphy ·
Australia's data center backlash is the social license problem that AI infrastructure spending cannot buy its way out of
Communities across Australian cities are organizing against large data centers over energy use, water consumption, noise, and planning opacity, surfacing a social license problem that hyperscalers and AI infrastructure investors have systematically underpriced. The resistance is an early signal of constraints appearing across Virginia, Ireland, and other data-center-dense markets, with real implications for compute availability, cloud pricing timelines, and the competitive advantage available to
Most Read
1
2
3
4
5
AI Briefing
5 most important AI updates after 8pm every day, curated for you