Business, Deals & Funding
Guardian AI
‘Poisoned’ AI: the ChatGPT shopping scams that lead to fake websites
The article discusses a scam where consumers are misled by ChatGPT into visiting fake shopping websites. When users ask ChatGPT for product recommendations from known brands like Russell & Bromley, the AI provides seemingly legitimate suggestions with prices and links. However, these links lead to counterfeit websites designed to look like official brand sites, resulting in consumers being ripped off. The phenomenon is described as 'poisoned' AI, where the AI tool's recommendations lend false credibility to fraudulent online stores, exploiting users' trust in AI-generated advice.
Why it matters
This article highlights a genuinely alarming and underreported vulnerability in the growing reliance on AI assistants for everyday tasks like shopping. The fact that users implicitly trust AI recommendations — and that bad actors can exploit this trust by manipulating AI outputs to direct people to fake websites — represents a significant consumer protection issue. It raises important questions about the responsibility of AI companies like OpenAI to verify the legitimacy of sources they link to…
TechCrunch AI
OpenAI unveils Lockdown Mode to protect sensitive data from prompt injection attacks
OpenAI has announced a new feature called Lockdown Mode designed to protect sensitive data from prompt injection attacks. When enabled, Lockdown Mode disables live web browsing (allowing only cached content), retrieval and display of web images, deep research, and agent mode. OpenAI acknowledges that ChatGPT could still be vulnerable to prompt injections even with Lockdown Mode active, as malicious instructions could appear in cached content or uploaded files. However, the feature aims to reduce the risk of sensitive data exfiltration. OpenAI says Lockdown Mode is not intended for all users but rather for people and organizations handling sensitive data who want stricter protections. The feature is currently rolling out to self-serve ChatGPT Business accounts and eligible personal accounts.
Why it matters
This is a significant and overdue acknowledgment by OpenAI that prompt injection is a serious, unsolved security problem — one that the AI industry has largely downplayed. The approach of disabling key capabilities like live browsing, image retrieval, deep research, and agent mode is essentially an admission that these features create attack surfaces that cannot yet be fully secured. It's a pragmatic, defense-in-depth strategy: if you can't prevent the injection, at least limit what the attacke…
Guardian AI
Suit filed against controversial planned Stratos datacenter project in Utah
Utah residents and the progressive non-profit Alliance for a Better Utah have filed a lawsuit against the planned Stratos AI datacenter project in Box Elder County, which is backed by Shark Tank investor Kevin O'Leary. The plaintiffs claim the project 'irrevocably' cuts off citizens' rights by not allowing sufficient public input. While the project's footprint has been reduced, concerns about health impacts remain.
Why it matters
This lawsuit highlights growing public resistance to the rapid expansion of AI datacenter infrastructure, particularly when local communities feel excluded from the decision-making process. The involvement of a celebrity investor like Kevin O'Leary adds visibility but also scrutiny. The residents' concerns about health impacts and lack of public input are legitimate governance issues that deserve proper adjudication. Large-scale datacenter projects can bring economic benefits but also impose si…
TechCrunch AI
What to expect from WWDC 2026: Siri’s highly anticipated revamp and Apple Intelligence updates
Apple's WWDC 2026, starting Monday, is expected to feature a major AI overhaul of Siri using Google's Gemini technology, making it more conversational and capable of multi-step tasks, along with a standalone Siri app to compete with ChatGPT and Claude. Other anticipated announcements include an AI agent integration with the App Store for task delegation, Visual Intelligence enhancements in the Camera app using Google Image Search, AI-powered photo editing with natural language commands, upgraded Image Playground with higher-quality image generation, and Apple Wallet updates including bill-splitting via receipt photos and digital pass creation from physical tickets.
Why it matters
This article is a fairly standard pre-event speculation roundup, aggregating leaks and rumors from sources like Bloomberg and The Information. The most striking detail is Apple's apparent deep integration of Google's Gemini technology into Siri, which represents a significant strategic shift — essentially acknowledging that Apple's in-house AI efforts haven't kept pace with competitors. The AI agent app store concept could be transformative if executed well, potentially creating a new ecosystem…
TechCrunch AI
Sriram Krishnan is leaving his role as White House AI advisor
Sriram Krishnan, a former tech executive and Andreessen Horowitz partner serving as the White House's senior AI policy advisor, is leaving the Trump administration at the end of June 2026. During his tenure, he helped shape the administration's AI Action Plan, which prioritized data center construction over regulation and safety. Krishnan worked closely with David Sacks, who earlier stepped down as AI and crypto czar. Krishnan plans to start a new outside institution that will continue to influence Trump's AI policy, focusing on energy, data centers, and AI benefits for Americans. His departure follows several AI-related executive orders, including ones challenging state-level AI regulations and a narrowed oversight order that was scaled back after industry pushback.
Why it matters
Krishnan's departure and immediate pivot to an outside institution designed to continue influencing Trump's AI policy raises concerns about the revolving door between Silicon Valley and government. The fact that the administration's AI approach has consistently prioritized industry interests — favoring data center construction over regulation, narrowing oversight after industry pushback, and even entertaining government equity stakes in AI companies — suggests that advisors like Krishnan have b…
Lenny's Newsletter
🧠 Community Wisdom: Bootstrapping vs. raising funding, building the roadmap of your vibe-coded app, AI agents and data integrity, your first project as an APM, and more
This is a paywalled edition of Lenny's Newsletter's 'Community Wisdom' series (edition 188), published on June 6, 2026. The topics teased in the title include: bootstrapping vs. raising funding, building a roadmap for a vibe-coded app, AI agents and data integrity, advice for first projects as an Associate Product Manager (APM), and more. The actual content is behind a paid subscriber wall and is not accessible, so no substantive details from the discussions are available.
Why it matters
Since the content is entirely paywalled and inaccessible, there's nothing to evaluate beyond the headline topics. The topics themselves sound relevant and timely for product managers and startup founders — particularly the intersection of vibe coding and product roadmapping, and the question of data integrity with AI agents. However, without being able to read the actual community discussions, I can't assess the quality or depth of the advice. The article is essentially a teaser with no usable…
NY Times
N.Y. Lawmakers Move to Pause Data Centers and Curb Surveillance Pricing
New York lawmakers are taking action to pause data center construction and curb surveillance pricing practices. The state also passed legislation initiating a two-year process to allow more frequent redrawing of congressional district lines. Meanwhile, efforts to contain plastics have stalled for a third consecutive year.
Why it matters
This article touches on several important policy areas. Pausing data center expansion reflects growing concerns about the energy and environmental impacts of AI infrastructure, which is a prudent step to assess before allowing unchecked growth. Curbing surveillance pricing — where companies use personal data to charge different customers different prices — is a welcome consumer protection measure. The redistricting legislation raises questions about potential gerrymandering risks, and the conti…
TechCrunch AI
The Trump administration might take an equity stake in OpenAI
The Trump administration is reportedly in discussions to take an equity stake in OpenAI, with President Trump stating he's been talking to AI companies about deals where 'the American people can benefit from the success of AI.' CNBC reported that some equity could seed a 'Public Wealth Fund' proposed by OpenAI, which would distribute proceeds directly to citizens. Bloomberg reports that CEO Sam Altman has been discussing the idea of government stakes in major AI companies since early 2025. This aligns with Trump's broader interest in government ownership of companies, such as the government's 10% stake in Intel. Senator Bernie Sanders has also proposed a 50% one-time stock tax on AI companies like OpenAI, Anthropic, and xAI. Former Trump AI czar David Sacks warned such moves could 'accelerate the corporate-government fusion,' while former Microsoft employee Dare Obasanjo suggested the g…
Why it matters
This is a deeply concerning development that blurs the lines between government and private enterprise in troubling ways. A government equity stake in OpenAI creates massive conflicts of interest: how can regulators objectively oversee AI safety and competition when the government is literally a shareholder benefiting from the company's success? The 'Public Wealth Fund' framing is clever marketing, but it essentially gives OpenAI political cover and a powerful government ally in exchange for eq…
From Reddit/HN/YC
- [Hacker News] The Economist argues Nike can't just do it any more, tracing how the brand lost its cultural grip.
- [Hacker News] Kyushu is a self-hostable WASM sandbox for JavaScript workers — isolate untrusted code without spinning up containers.
- [Hacker News] WinkTerm is an AI that shares the same PTY as you in the terminal, watching and acting alongside your shell session.
- [Hacker News] New paper examines how AI agents spend your money — tracking the real-world cost behavior of autonomous tool-using systems.
- [Hacker News] Wired reviewer says the GroWell Cap gave him hair for the first time in 15 years — laser cap therapy apparently works.
- [Hacker News] Meta will give staff 30-minute breaks from keylogging in its employee monitoring scheme. Generous.