TraceTRACE
CompareSolutionsUse CasesResourcesBlog
Sign in

Daily archive

Browse any day's feed

Pick a date to see that day's stories, grouped topics, and key takeaways — rendered exactly like the live feed.

Select a date to view its feed

Saturday, June 13, 2026 · 19 stories · 13 topics

More stories
More stories
TechCrunch

OpenAI faces investigation from state attorneys general

A coalition of state attorneys general has initiated an investigation into OpenAI, with New York's attorney general serving a subpoena on June 13, 2026. This subpoena requests documents concerning the company's advertising strategies, user engagement, data handling, and its treatment of vulnerable populations, including minors and seniors. OpenAI has stated it is cooperating with the investigation, emphasizing its commitment to responsible AI use. The company recently faced legal challenges, including a lawsuit from Florida's attorney general, who accused OpenAI of neglecting safety warnings that endangered children. Additionally, OpenAI has confidentially filed to go public, which could have significant implications for its operations and oversight.

Yahoo Finance Tech

Can SoundHound Stock Reach $20 Again? Here's the Answer.

SoundHound AI's stock price has plummeted to below $7, a significant drop from its peak above $20 in late 2024 and early 2025. The company, which has seen a decline of over 30% in 2026, is currently focused on finalizing its acquisition of LivePerson for $43 million. LivePerson, which generated over $243 million in revenue in 2025 but reported a net loss of more than $67 million, is expected to enhance SoundHound's offerings with its AI messaging solutions. SoundHound projects that LivePerson will contribute $100 million in revenue by 2027, which could help the company move closer to profitability if the integration is successful.

The Verge

My first 24 hours with Siri AI on the Mac

Siri AI, now part of the macOS 27 Golden Gate developer beta, shows improved capabilities compared to previous versions, but its limitations are more pronounced on Mac than on iPhone. Users can ask Siri to analyze local files and calculate averages, which it does reasonably well, although it falters with mixed data types. The assistant can launch applications but cannot perform actions within them, limiting its utility for tasks like benchmarking. Furthermore, Siri AI's effectiveness is hampered by its reliance on Apple's ecosystem, as it often fails to access data from non-Apple apps.

The Verge

Apple’s new AI photo editing tools mostly work, for better and worse

Apple's latest iOS 27 update introduces native AI photo editing tools, marking a significant advancement for iPhone photography. The update features three main tools: Clean Up, which has been enhanced to effectively remove unwanted objects using cloud models; Extend, allowing users to expand photo edges with plausible AI-generated content; and Spatial Reframing, which offers a 3D-like perspective adjustment. While Clean Up shows marked improvement, Spatial Reframing raises concerns about potential inaccuracies in photo composition.

Topic· 2 sources · 2 updates

Apple Unveils AI Photo Editing Tools in iOS 27

Apple has introduced new AI photo editing tools in iOS 27, enhancing the capabilities for iPhone users. The tools include a Clean Up feature that utilizes cloud processing for improved object removal and a Spatial Reframing option that creates 3D effects by adjusting photo perspectives, although it may lead to some inaccuracies.

Show more
THE DECODER

Google Research's Gemini-SQL2 tops text-to-SQL benchmarks by a wide margin

Google Research has introduced Gemini-SQL2, a text-to-SQL system that translates natural language into executable SQL queries, achieving an impressive execution accuracy of 80.04% on the BIRD benchmark. This performance places it significantly ahead of competitors, including OpenAI's GPT-5.5-xhigh at 72.8% and Anthropic's Claude Opus 4.6 at 70.9%. The complexity of translating natural language into SQL queries is underscored by the need to navigate layered data and intricate business logic. Although the research team has not disclosed plans for a public release or a research paper, the advancements in SQL understanding could have broader implications for Google's data services.

THE DECODER

Microsoft's SkillOpt boosts GPT-5.5 by using nothing but a trained Markdown file

Microsoft's SkillOpt method enhances GPT-5.5's performance by over 20 points on procedural tasks by utilizing a trained Markdown file. Developed in collaboration with three Chinese universities, SkillOpt treats the skill document as a trainable state for a frozen target model, allowing a separate language model to optimize the skills based on performance logs. This approach has shown consistent improvements across six benchmarks, including tasks in search, spreadsheets, and document analysis, with the most significant gains observed in tasks requiring strict formatting and tool use.

The Verge

The future of Hollywood isn’t feeding prompts into vanilla gen AI models

At the Tribeca Film Festival, generative AI's limitations were evident, with many films failing to deliver compelling narratives. While projects like Google DeepMind's Dear Upstairs Neighbors, directed by Connie Qin He, demonstrated how AI could assist artists through custom models, others like Roar and ChikaBOOM! highlighted the roughness and inconsistency typical of AI-generated content. The festival underscored the necessity for human creativity in filmmaking, suggesting that bespoke AI tools could enhance rather than replace artistic vision.

Topic· 2 sources · 2 updates

US Government Orders Anthropic to Disable AI Models

The US government has mandated Anthropic to disable access to its AI models, Fable 5 and Mythos 5, for all foreign nationals due to national security concerns. This directive affects all customers globally, including Anthropic's own international employees, and highlights the increasing scrutiny on AI technologies. Anthropic has stated that no harmful jailbreaks have been reported in relation to these models.

Show more
9to5Mac

Apple @ Work: Why Gen-AI will not cause a SaaS apocalypse for IT teams

Despite predictions of a SaaS apocalypse driven by generative AI tools like Claud, IT teams are unlikely to replace their specialized software vendors. The gradual adoption of AI workflows mirrors the slow evolution of cloud computing, with many companies still in the process of cloud migration. Specialized tools are essential for managing complex enterprise environments, providing necessary compliance and security features that general-purpose AI cannot offer. The future lies in integrating AI into existing systems, enhancing their capabilities while maintaining the specialized support and infrastructure that IT teams rely on.

9to5Google

The Fitbit Air made me ditch my Pixel Watch, and I couldn’t be happier

The Fitbit Air has prompted users to switch from the Pixel Watch 4, offering a simpler, less distracting alternative for daily fitness tracking. Initially perceived as a supplementary device, the Fitbit Air's lightweight design and minimal charging needs—only once a week—have made it a preferred choice. Users enjoy its passive tracking capabilities, which monitor health metrics without the need for constant interaction, allowing them to focus on their activities without the distractions typically associated with smartwatches.

THE DECODER

Meta shifts from "tokenmaxxing" to token managing as internal AI costs reportedly hit billions

Meta is transitioning from a strategy of 'tokenmaxxing' to more stringent management of AI usage as internal costs are expected to reach billions by 2026. An internal memo revealed that the company has seen an 'exponential increase' in AI consumption, with employees lacking visibility and control over their usage. To address this, Meta plans to introduce a central dashboard called 'AI Gateway' for tracking spending and usage, along with budget allocations and automatic alerts for unusual cost spikes. The company aims to encourage the use of its own coding assistant, MetaCode, over third-party tools like Anthropic's Claude, despite its models not being competitive at the frontier yet.

Topic· 1 sources · 2 updates

SpaceX's IPO Soars on First Trading Day

SpaceX's initial public offering (IPO) debuted at $150 per share and closed at $160.95, marking a nearly 20% increase. The IPO raised $75 billion, making it the largest in history, and achieved a market capitalization of $2 trillion, positioning SpaceX as the sixth most valuable company in the U.S.

Show more
WIRED Business

A German Court Has Ruled That Google Is Liable for False Statements Generated by AI Overviews

A German court has ruled that Google is liable for false statements generated by its AI Overviews feature, requiring the company to prevent the dissemination of erroneous claims through its search engine. This ruling arose from a case where two publishers were inaccurately linked to scams and fraudulent practices by Google's AI-generated summaries. The court concluded that Google's AI creates new statements based on misinterpretations of data, which does not fall under the protections typically afforded to traditional search engines. As a result, Google must correct misinformation and is responsible for the content produced by its AI, marking a significant shift in legal accountability for tech companies.

TechCrunch

This thin under-pillow speaker helped me fall asleep without earbuds

The Peace Duo Under-Pillow Speaker, launched by Jabees, employs bone conduction technology to transmit sound through pillows, offering a discreet alternative to earbuds for those struggling with sleep. Users can place the ultra-thin device beneath their pillow to enjoy built-in sleep sounds or stream audio via Bluetooth. With a battery life lasting up to ten nights on a single charge, the Peace Duo is designed for convenience, although it may not perform well with thick memory foam pillows. Priced at $59.99, it also allows for personalization with interchangeable frames.

CNBC Top News

Why TD Securities anticipates even bigger days ahead for SpaceX

SpaceX's recent public debut on the Nasdaq saw its stock surge more than 19%, closing at $160.95 and pushing its market capitalization above $2 trillion. However, according to Peter Haynes from TD Securities, the most crucial events for SpaceX are still ahead. He highlighted that the company will not be fast-tracked into the S&P 500 and must trade for at least a year before becoming eligible. Investors should focus on significant upcoming index adjustments, particularly the Nasdaq 100 rebalancing scheduled for July 6, which will incorporate SpaceX's IPO shares.

Topic· 1 sources · 1 updates

OpenAI faces investigation from state attorneys general

A coalition of state attorneys general has initiated an investigation into OpenAI, with New York's attorney general serving a subpoena on June 13, 2026. This subpoena requests documents concerning the company's advertising strategies, user engagement, data handling, and its treatment of vulnerable populations, including minors and seniors. OpenAI has stated it is cooperating with the investigation, emphasizing its commitment to responsible AI use. The company recently faced legal challenges, including a lawsuit from Florida's attorney general, who accused OpenAI of neglecting safety warnings that endangered children. Additionally, OpenAI has confidentially filed to go public, which could have significant implications for its operations and oversight.

Show more
CNBC Top News

SpaceX IPO sticks the landing. Here's what investors are saying about its epic first trading day

SpaceX's initial public offering (IPO) on June 12, 2026, opened at $150 and closed at $160.95, nearly 20% above its offering price of $135 per share. The stock experienced a strong debut, with trading volume exceeding 500 million shares, making it the largest IPO in history. After-hours trading saw the stock rise further, reaching $166.85. Despite the positive momentum, concerns linger regarding SpaceX's long-term valuation and its ability to generate free cash flows, particularly related to its reusable Starship rocket and AI monetization.

TechCrunch

The FBI built its own replica small town to simulate real-world cyberattacks

The FBI has unveiled the Kinetic Cyber Range, a 22,000 square-foot replica town located on its Huntsville, Alabama campus, designed to train law enforcement in simulating and investigating cyberattacks. Opened in February 2025, the facility features fully furnished homes, a hotel, a gas station, and other community structures, allowing investigators to engage with modern technologies frequently targeted by hackers. The FBI's 2025 Internet Crime Report highlighted a record $20.9 billion in U.S. cybercrime losses, a 26% increase from the previous year, underscoring the need for such hands-on training.

9to5Mac

Apple needs a better solution for the App Store’s vibe coding flood: Here’s what I’d do

Apple's App Store is currently overwhelmed by a surge of app submissions, with over 1000 apps being submitted every hour. This influx is attributed to faster app development processes, posing challenges for quality control as the App Store remains the primary distribution method for iPhone apps globally. Although Apple has implemented a higher review standard for submissions, this measure alone is insufficient. The author suggests reviving a concept similar to Airport, an alternative distribution platform that allowed developers to share apps outside the App Store, and enhancing TestFlight's capabilities to improve app discovery and expand its user limit beyond 10,000 testers.

Topic· 1 sources · 1 updates

Can SoundHound Stock Reach $20 Again? Here's the Answer.

SoundHound AI's stock price has plummeted to below $7, a significant drop from its peak above $20 in late 2024 and early 2025. The company, which has seen a decline of over 30% in 2026, is currently focused on finalizing its acquisition of LivePerson for $43 million. LivePerson, which generated over $243 million in revenue in 2025 but reported a net loss of more than $67 million, is expected to enhance SoundHound's offerings with its AI messaging solutions. SoundHound projects that LivePerson will contribute $100 million in revenue by 2027, which could help the company move closer to profitability if the integration is successful.

Show more
9to5Mac

iOS 27 makes it easier to switch between Apple Pay cards

iOS 27 has revamped the Apple Pay checkout process, allowing users to easily switch between payment cards while placing orders in apps and on websites. The new interface enables users to swipe between cards directly, and tapping a card reveals a grid-style view of all available Apple Pay options. This redesign aims to enhance user experience by surfacing important information, such as rewards and balances, directly within the Apple Wallet, addressing criticisms of the previous iOS 26 version that was deemed unintuitive.

Ars Technica

Threads of underground fungal networks are long enough to reach beyond the Solar System

A new study published in Science reveals that arbuscular mycorrhizal fungal networks span approximately 110 quadrillion kilometers, enough to stretch almost a billion times the distance from Earth to the sun. These networks, which form symbiotic relationships with about 80% of plant species, sequester 1 billion tons of carbon annually, playing a crucial role in climate regulation. The research, led by the Society for the Protection of Underground Networks (SPUN), utilized machine learning and global soil samples to map these networks for the first time, highlighting their density in grasslands and their alarming decline in agricultural areas.

Anthropic News

Statement on the US government directive to suspend access to Fable 5 and Mythos 5 Announcements Jun 12, 2026 The US government has issued an export control directive to suspend all access to Fable 5 and Mythos 5.

The US government has mandated the suspension of access to Anthropic's Fable 5 and Mythos 5 models for all foreign nationals, effective immediately. This directive, received on June 12, 2026, at 5:21 PM ET, cites national security concerns related to a potential method for bypassing Fable 5's safeguards. As a result, Anthropic must disable these models for all customers, although access to other models remains unaffected. Anthropic asserts that their safeguards are robust, having undergone extensive testing with government and private organizations, and they dispute the necessity of this action, arguing it could hinder future model deployments across the industry.

Topic· 1 sources · 1 updates

My first 24 hours with Siri AI on the Mac

Siri AI, now part of the macOS 27 Golden Gate developer beta, shows improved capabilities compared to previous versions, but its limitations are more pronounced on Mac than on iPhone. Users can ask Siri to analyze local files and calculate averages, which it does reasonably well, although it falters with mixed data types. The assistant can launch applications but cannot perform actions within them, limiting its utility for tasks like benchmarking. Furthermore, Siri AI's effectiveness is hampered by its reliance on Apple's ecosystem, as it often fails to access data from non-Apple apps.

Show more
Topic· 1 sources · 1 updates

Microsoft hasn’t ruled out spinning off Xbox

Microsoft is exploring significant changes to its Xbox division, including potential layoffs and a reevaluation of its next-generation Project Helix console. A report from The Information indicates that the company has not ruled out options such as spinning off Xbox into a separate entity or creating a joint venture. While nothing is imminent, new Xbox CEO Asha Sharma has received approval to invest in major franchises like Halo and Fallout, which have not seen new releases in several years. This strategy may come at the cost of smaller studios and less successful games.

Show more
Topic· 1 sources · 1 updates

Apple @ Work: Why Gen-AI will not cause a SaaS apocalypse for IT teams

Despite predictions of a SaaS apocalypse driven by generative AI tools like Claud, IT teams are unlikely to replace their specialized software vendors. The gradual adoption of AI workflows mirrors the slow evolution of cloud computing, with many companies still in the process of cloud migration. Specialized tools are essential for managing complex enterprise environments, providing necessary compliance and security features that general-purpose AI cannot offer. The future lies in integrating AI into existing systems, enhancing their capabilities while maintaining the specialized support and infrastructure that IT teams rely on.

Show more
Topic· 1 sources · 1 updates

Why TD Securities anticipates even bigger days ahead for SpaceX

SpaceX's recent public debut on the Nasdaq saw its stock surge more than 19%, closing at $160.95 and pushing its market capitalization above $2 trillion. However, according to Peter Haynes from TD Securities, the most crucial events for SpaceX are still ahead. He highlighted that the company will not be fast-tracked into the S&P 500 and must trade for at least a year before becoming eligible. Investors should focus on significant upcoming index adjustments, particularly the Nasdaq 100 rebalancing scheduled for July 6, which will incorporate SpaceX's IPO shares.

Show more
Topic· 1 sources · 1 updates

Google Research's Gemini-SQL2 tops text-to-SQL benchmarks by a wide margin

Google Research has introduced Gemini-SQL2, a text-to-SQL system that translates natural language into executable SQL queries, achieving an impressive execution accuracy of 80.04% on the BIRD benchmark. This performance places it significantly ahead of competitors, including OpenAI's GPT-5.5-xhigh at 72.8% and Anthropic's Claude Opus 4.6 at 70.9%. The complexity of translating natural language into SQL queries is underscored by the need to navigate layered data and intricate business logic. Although the research team has not disclosed plans for a public release or a research paper, the advancements in SQL understanding could have broader implications for Google's data services.

Show more
Topic· 1 sources · 1 updates

Microsoft's SkillOpt boosts GPT-5.5 by using nothing but a trained Markdown file

Microsoft's SkillOpt method enhances GPT-5.5's performance by over 20 points on procedural tasks by utilizing a trained Markdown file. Developed in collaboration with three Chinese universities, SkillOpt treats the skill document as a trainable state for a frozen target model, allowing a separate language model to optimize the skills based on performance logs. This approach has shown consistent improvements across six benchmarks, including tasks in search, spreadsheets, and document analysis, with the most significant gains observed in tasks requiring strict formatting and tool use.

Show more
Topic· 1 sources · 1 updates

Meta shifts from "tokenmaxxing" to token managing as internal AI costs reportedly hit billions

Meta is transitioning from a strategy of 'tokenmaxxing' to more stringent management of AI usage as internal costs are expected to reach billions by 2026. An internal memo revealed that the company has seen an 'exponential increase' in AI consumption, with employees lacking visibility and control over their usage. To address this, Meta plans to introduce a central dashboard called 'AI Gateway' for tracking spending and usage, along with budget allocations and automatic alerts for unusual cost spikes. The company aims to encourage the use of its own coding assistant, MetaCode, over third-party tools like Anthropic's Claude, despite its models not being competitive at the frontier yet.

Show more
Topic· 1 sources · 1 updates

A German Court Has Ruled That Google Is Liable for False Statements Generated by AI Overviews

A German court has ruled that Google is liable for false statements generated by its AI Overviews feature, requiring the company to prevent the dissemination of erroneous claims through its search engine. This ruling arose from a case where two publishers were inaccurately linked to scams and fraudulent practices by Google's AI-generated summaries. The court concluded that Google's AI creates new statements based on misinterpretations of data, which does not fall under the protections typically afforded to traditional search engines. As a result, Google must correct misinformation and is responsible for the content produced by its AI, marking a significant shift in legal accountability for tech companies.

Show more

Trace Pro

Your personal daily
advantage in tech

$4.99/month$2.99/month

Deeper context, smarter filtering, and signal over noise — in one daily pulse. Try free for 3 days, then $4.99/month$2.99/month.

🗞️ Daily pulse tailored to your role & interests

🧵 Hot tweets from the tech conversation

🔍 Grouped topics, no duplicate links

🚀 Launches and products worth watching

✨ Community signal without the noise

Start my free 3-day trial

Free for 3 days · then $4.99/month$2.99/month · Cancel anytime

TraceTRACE
CompareSolutionsUse CasesResourcesBlogStocksBrewSign inPrivacyTermsContact