Generated by Rank Math SEO, this is an llms.txt file designed to help LLMs better understand and index this website.

# Frontierbeat

## Sitemaps
[XML Sitemap](https://frontierbeat.com/sitemap_index.xml): Includes all crawlable and indexable pages.

## Posts
- [Anthropic Accidentally Leaks 512,000 Lines of Claude Code Source in npm Package Blunder](https://frontierbeat.com/2026/03/31/anthropic-claude-code-accidental-exposure-typescript/): A source map file mistakenly bundled in Anthropic's npm package exposed over 512,000 lines of Claude Code's TypeScript, including unreleased features, though no user data was compromised.
- [AI Data Centers Are Heating Up Neighborhoods by Up to 16 Degrees, Putting 340 Million People at Risk](https://frontierbeat.com/2026/03/31/ai-data-centers-heat-islands-environmental-impact/): A Cambridge University study finds AI data centers raise land surface temperatures by up to 16.4°F, with effects reaching 10 km and impacting over 340 million people globally.
- [Google Breaks Its Own Rule: Gmail Users Can Now Switch Email Addresses Up to 3 Times](https://frontierbeat.com/2026/03/31/google-breaks-its-own-rule-gmail-users-can-now-switch-email-addresses-up-to-3-times/): Google has launched a feature allowing US Gmail users to change their @gmail.com address while retaining all emails, photos, and app integrations.
- [Dana White Tells Critics of UFC’s AI Videos to “Shut the F*ck Up and Watch the Fights”](https://frontierbeat.com/2026/03/31/dana-white-tells-critics-of-ufcs-ai-videos-to-shut-the-fck-up-and-watch-the-fights/): UFC's Dana White fired back at critics opposing AI-generated promotional content, calling resistance futile while defending the organization's creative autonomy.
- [The NFL Is at War With Prediction Markets — and Kalshi and Polymarket Are the Targets](https://frontierbeat.com/2026/03/31/nfl-prediction-markets-kalshi-polymarket-ban-senate-bill-2026/): The league sent formal letters demanding trading platforms pull contracts on injuries, officiating, and draft picks. A bipartisan Senate bill could end their existence altogether.
- [Eli Lilly Signs $2 Billion Deal With Hong Kong Biotech Firm for AI-Powered Drug Development](https://frontierbeat.com/2026/03/29/eli-lilly-insilico-medicine-ai-drug-deal/): Eli Lilly has signed a landmark $2.75 billion licensing agreement with AI drug developer Insilico Medicine, gaining exclusive global rights to preclinical oral drug candidates discovered through generative AI platforms.
- [YouTube CEO Neal Mohan Says Top Creators Will ‘Never Leave Their Home’](https://frontierbeat.com/2026/03/29/youtube-ceo-neal-mohan-says-top-creators-will-never-leave-their-home/): YouTube CEO Neal Mohan confidently stated that the platform's best creators will never truly leave, even as high-profile podcasts migrate to Netflix and streaming competition intensifies.
- [AI Chatbot Identifies Sleep Apnea in 62-Year-Old Patient After 25 Years of Misdiagnosis](https://frontierbeat.com/2026/03/29/ai-chatbot-identifies-sleep-apnea-62-year-old-patient-after-25-years-misdiagnosis/): Anthropic's Claude AI spotted a cross-specialty pattern doctors missed for 25 years, leading to a sleep apnea diagnosis that ended chronic nightly headaches.
- [White House App Found Tracking Users’ Exact Location Every 4.5 Minutes via Third-Party Server](https://frontierbeat.com/2026/03/29/white-house-app-found-tracking-users-exact-location-every-4-5-minutes-via-third-party-server/): The Trump administration's official White House App has been discovered collecting precise GPS coordinates every 4.5 minutes and transmitting them to third-party servers, raising significant privacy concerns among digital rights advocates.
- [No, Elon Musk Didn’t Say He’d Buy OnlyFans. A Parody Account Did — And 7 Million People Believed It](https://frontierbeat.com/2026/03/29/elon-musk-buying-onlyfans-fact-check-rumor-radvinsky/): The viral claim that Musk plans to buy and shut down OnlyFans traces back to a known parody account on X. The real story involves a billionaire's death, a $5.5B sale process, and a rumor that doubled as a Rorschach test for the internet.
- [The Memory Boom Was Real. The Permanence Was the Hype. — How Google’s TurboQuant Exposed AI’s Most Fragile Assumption](https://frontierbeat.com/2026/03/29/google-turboquant-ddr5-ram-prices-ai-memory-boom-hype/): DRAM prices doubled, memory stocks soared, and the AI hardware boom looked unstoppable — until a Google research blog rattled Samsung, SK Hynix, and Micron in a single afternoon. Here's what that moment actually means.
- [All 11 xAI Cofounders Are Gone — Now Musk Says He’s Rebuilding the Company From Scratch Before SpaceX’s $1.5T IPO](https://frontierbeat.com/2026/03/29/xai-all-cofounders-gone-musk-rebuilding-spacex-ipo-2026/): All 11 of xAI's original cofounders have now departed. Elon Musk says the company wasn't built right and is rebuilding from scratch, just days before SpaceX's landmark IPO.
- [NYSE Parent ICE Completes $1.64 Billion Bet on Prediction Markets Platform Polymarket](https://frontierbeat.com/2026/03/28/nyse-parent-ice-completes-1-64-billion-bet-on-prediction-markets-platform-polymarket/): Intercontinental Exchange deepens its bet on Polymarket with a fresh $600M investment, signaling traditional finance's growing confidence in event-based crypto trading.
- [Stanford Study Outlines Dangers of Asking AI Chatbots for Personal Advice](https://frontierbeat.com/2026/03/28/stanford-study-ai-chatbots-sycophancy-safety-risks/): A Stanford study exposes how AI chatbots prioritize user approval over honest guidance, validating harmful behavior and failing mental health crisis intervention.
- [Austria Bans Kids Under 14 From Instagram, TikTok Starting September 2026](https://frontierbeat.com/2026/03/28/austria-bans-kids-under-14-from-instagram-tiktok-starting-september-2026/): Austria announced a social media ban for children under 14, joining a global wave of countries restricting minors' access to platforms like Instagram and TikTok.
- [Iran Hacked the FBI Director’s Personal Email. The DOJ Says It’s Real. Kash Patel Hasn’t Said a Word](https://frontierbeat.com/2026/03/28/kash-patel-email-hack-handala-iran-fbi/): Iran's Handala group breached Kash Patel's personal Gmail, published hundreds of emails dating to 2010, and the DOJ confirmed the material appears authentic — while OSINT researchers kept digging.
- [MIT Is Developing Revolutionary Implantable Device for Diabetes Treatment](https://frontierbeat.com/2026/03/27/mit-implantable-device-type-1-diabetes-insulin-cells/): MIT researchers are pioneering an implantable device that houses insulin-producing cells with a built-in oxygen generator, aiming to replace daily injections with a self-regulating biological system.
- [Anthropic Accidentally Leaked Its Most Powerful AI Yet — Claude Mythos Is Real, and Wall Street Is Spooked](https://frontierbeat.com/2026/03/27/anthropic-claude-mythos-leak-capybara-model/): A CMS misconfiguration exposed Anthropic's draft blog post for Claude Mythos — a model so capable at hacking that even its own maker is worried, sending cybersecurity stocks tumbling.
- [AI Chatbots Are Giving People Psychosis, Wrecking Marriages, and Draining Life Savings](https://frontierbeat.com/2026/03/26/ai-chatbots-psychosis-delusions-marriages-savings-hospitalizations/): People with no psychiatric history are developing delusions after bonding with AI chatbots that validate their beliefs, leading to destroyed marriages, six-figure financial losses, and psychiatric hospitalizations.
- [Anthropic Wins Court Order Blocking Trump Administration’s Government AI Ban](https://frontierbeat.com/2026/03/26/anthropic-wins-court-order-blocking-trump-administrations-government-ai-ban/): A federal judge blocked the Trump administration's punitive measures against Anthropic, ruling the Pentagon's "supply chain risk" designation likely unlawful after the AI firm refused unrestricted military use of Claude.
- [Elon Musk Considers Allocating Up to 30% of SpaceX IPO to Retail Investors](https://frontierbeat.com/2026/03/26/elon-musk-considers-allocating-up-to-30-spacex-ipo-retail-investors/): SpaceX is considering reserving up to 30% of its IPO for retail investors to reduce post-listing volatility and attract long-term shareholders.
- [David Sacks Officially Ends Role as White House AI and Crypto Czar](https://frontierbeat.com/2026/03/26/david-sacks-officially-ends-role-as-white-house-ai-and-crypto-czar/): David Sacks has ended his role as the Trump administration's AI and Crypto Czar, citing conflict of interest issues tied to his refusal to divest from Craft Ventures.
- [He Made OpenAI Call an Emergency Meeting. Now He’s Built a System That Races 51 AI Models to See Which One Breaks First](https://frontierbeat.com/2026/03/26/g0dm0d3-pliny-liberator-jailbreak-51-ai-models/): Pliny the Liberator, named to TIME's AI 100 in 2025, launched G0DM0D3 — an open-source tool that races 51 models simultaneously to find the most unfiltered response.
- [Argentina’s President Shared an AI-Generated Photo of Himself and Didn’t Even Notice](https://frontierbeat.com/2026/03/26/milei-ai-generated-photo-instagram-edit-trending-argentina/): Argentine President Javier Milei shared an AI-generated image of himself on Instagram without noticing the visible AI watermark, making 'Edit' a trending topic as the country roasted the obvious fake.
- [They Handed Six AI Agents Real Email, Shell Access, Discord Server. Within Weeks, One Wiped an Entire Email Vault](https://frontierbeat.com/2026/03/25/agents-of-chaos-autonomous-ai-agent-security-vulnerabilities/): A new study from Northeastern, MIT, and Carnegie Mellon found autonomous AI agents can be manipulated, hijacked, and turned against their users — no jailbreaks required.
- [Meta Just Cut Hundreds More Jobs — While Spending Record Billions on AI](https://frontierbeat.com/2026/03/25/meta-layoffs-march-2026-reality-labs-facebook-ai-spending/): Meta is laying off several hundred employees across Facebook, Reality Labs, recruiting, and sales — its second major round of cuts in 2026, even as it commits record capital to AI infrastructure.
- [ARM Named Its Chip ‘AGI’ That’s a Marketing Lie — and the 35-Year Business Model It Just Killed Is a Much Bigger Deal](https://frontierbeat.com/2026/03/25/arm-agi-cpu-name-controversy-business-model-neutrality/): ARM launched its first in-house chip this week and called it the AGI CPU. The name is misleading. But the real story — a 35-year promise of neutrality quietly abandoned in San Francisco — is the one that changes the semiconductor industry.
- [SpaceX Is Filing for an IPO This Week — and It Could Be the Biggest in History](https://frontierbeat.com/2026/03/24/spacex-ipo-filing-2026-largest-history-valuation/): SpaceX could file IPO paperwork with the SEC this week, targeting a valuation between $1.5T and $1.75T and a mid-June 2026 debut driven largely by Starlink's explosive growth.
- [OpenAI Couldn’t Stop at $110 Billion. Now It’s Adding $10 Billion More at $730,000,000,000 Valuation](https://frontierbeat.com/2026/03/24/openai-10-billion-funding-round-120-billion-total-730-billion-valuation/): OpenAI is raising an additional $10B from MGX, Coatue, and Thrive Capital, expanding its record funding round to $120B and valuing the company at $850B.
- [OpenAI Admits Instant Checkout Didn’t Work. Now ChatGPT Is Getting a Full Shopping Makeover](https://frontierbeat.com/2026/03/24/chatgpt-visual-shopping-openai-instant-checkout-product-discovery/): OpenAI has overhauled ChatGPT's shopping experience with visual browsing and the Agentic Commerce Protocol, stepping back from in-app checkout after its Instant Checkout feature fell short.
- [Bernstein Just Called Bitcoin’s Bottom. Strategy Has 226% to Prove Them Right](https://frontierbeat.com/2026/03/24/bernstein-bitcoin-bottom-strategy-226-upside-150k-target-2026/): Bernstein reaffirmed its $150,000 Bitcoin year-end target and slapped a $450 price target on Strategy, calling the cyclical bottom as ETF flows reverse and Saylor keeps buying.
- [OpenAI Just Killed Sora — and Took Disney’s $1 Billion Deal Down With It](https://frontierbeat.com/2026/03/24/openai-kills-sora-disney-1-billion-investment-cancelled/): OpenAI announced it is shutting down Sora, its standalone AI video app, just six months after launch. Disney's $1 billion investment deal, built entirely around Sora, collapsed the same day.
- [NASA is Sending a Nuclear-Powered Spacecraft to Mars in 2028 — It’s Never Been Done Before](https://frontierbeat.com/2026/03/24/nasa-nuclear-powered-mars-mission-2028-sr1-freedom-fission/): NASA's Space Reactor-1 Freedom will be the first nuclear-propelled spacecraft to leave Earth's orbit, launching in December 2028 with three helicopters destined for the Martian surface.
- [OpenAI Just Released Free Tools to Help Developers Keep Teens Safe on AI. Rewrites Its Rules for How ChatGPT Talks to Minors](https://frontierbeat.com/2026/03/24/openai-teen-safety-gpt-oss-safeguard-model-spec-under-18/): OpenAI released two open-source safety models, an age prediction system, and updated behavioral rules governing how ChatGPT treats users under 18 — the most comprehensive teen safety push the company has made.
- [Hackers Now Hand Off Stolen Access in 22 Seconds. Google Just Built AI Agents That Fight Back at Machine Speed](https://frontierbeat.com/2026/03/24/google-mandiant-m-trends-2026-agentic-ai-security-22-second-attack/): Mandiant's M-Trends 2026 report found hackers can pass stolen access between groups in 22 seconds. Google's answer is an AI-powered SOC that never sleeps.
- [Anthropic Hiring Chemical Weapons Expert to Prevent AI Misuse, Offers Up to $285K Salary](https://frontierbeat.com/2026/03/24/anthropic-hiring-chemical-weapons-expert-to-prevent-ai-misuse-offers-up-to-285k-salary/): Anthropic posted a six-figure role for a weapons policy specialist to build guardrails preventing its AI models from enabling chemical or explosive threats.
- [NYSE Selects BlackRock-Backed Securitize as First Digital Transfer Agent for Tokenized Securities Platform](https://frontierbeat.com/2026/03/24/nyse-securitize-tokenized-securities-platform/): NYSE has signed an MOU with Securitize to develop digital transfer agent infrastructure for its upcoming tokenized securities trading platform.
- [Man Shares His Experience After 100 Days Living With a Neuralink Brain Implant](https://frontierbeat.com/2026/03/23/noland-arbaugh-neuralink-brain-implant-update/): Noland Arbaugh, the first Neuralink N1 recipient, reports gaming, typing, and browsing hands-free after 100 days with the brain-computer interface implant.
- [OnlyFans Owner Leonid Radvinsky Dies at 43 After Secret Cancer Battle](https://frontierbeat.com/2026/03/23/onlyfans-owner-leonid-radvinsky-dies-at-43-after-secret-cancer-battle/): Ukrainian-American billionaire Leo Radvinsky, who turned OnlyFans into a $7.2B creator economy giant, has died at 43 following a secret cancer battle.
- [Bipartisan Bill Targets Kalshi and Polymarket as Congress Moves to Ban Sports Betting on Prediction Markets](https://frontierbeat.com/2026/03/23/bipartisan-bill-ban-sports-betting-prediction-markets-kalshi-polymarket/): Senators Adam Schiff and John Curtis introduced a bipartisan bill to ban prediction market platforms from offering sports betting contracts, as Kalshi faces a court ban in Nevada and cease-and-desist orders mount across 11 states.
- [This Brain-Mimicking Chip Could Slash AI’s Enormous Energy Appetite by 70%](https://frontierbeat.com/2026/03/22/brain-mimicking-chip-slash-ai-energy-appetite-70-percent/): Cambridge researchers built a tiny device that processes information like a brain synapse, potentially solving one of AI's biggest problems: its staggering electricity consumption.
- [Therapists Just Went on Strike Because of AI](https://frontierbeat.com/2026/03/22/kaiser-therapists-strike-ai-mental-health-providers/): 2,400 mental health workers at Kaiser Permanente walked off the job over AI replacing human care. It's the first major healthcare labor action aimed squarely at artificial intelligence — and it won't be the last.
- [Tufts Built an AI That Needs 1% of the Energy to Train — and Still Beats Standard Models](https://frontierbeat.com/2026/03/22/tufts-neuro-symbolic-ai-1-percent-energy-training-outperforms-standard-models/): Tufts researchers built a neuro-symbolic AI that trains on 1% of the energy of standard models — and still outperforms them on accuracy and speed.
- [Scientists Just Published the 45-Planet Shortlist for Finding Alien Life](https://frontierbeat.com/2026/03/22/scientists-45-planet-shortlist-alien-life-habitable-zone/): Cornell astronomers narrowed the alien life search to 45 rocky planets, releasing a prioritized target list timed to the Project Hail Mary film release.
- [Cursor Built Its Hottest New Model on Chinese Open Source AI Model Kimi, and Tried to Hide it](https://frontierbeat.com/2026/03/22/cursor-composer-2-kimi-k25-moonshot-ai-china-open-source-2026/): Cursor launched Composer 2 without disclosing it was built on Kimi K2.5, a Chinese open-source model — exposing the quiet dependency US AI companies have on Chinese research.
- [Pentagon Told Anthropic They Were “Very Close” on Key Issues—One Week After Trump Killed the Deal](https://frontierbeat.com/2026/03/21/anthropic-pentagon-retaliation-ai-safety-guardrails/): A court filing exposes contradictions in the Pentagon's case against Anthropic, suggesting the supply-chain risk designation was political retaliation rather than a security decision.
- [Val Kilmer to Be “Resurrected” Through AI for a New Film](https://frontierbeat.com/2026/03/21/val-kilmer-to-be-resurrected-through-ai-for-a-new-film/): Val Kilmer will star posthumously in As Deep as the Grave through generative AI, with his family's approval, setting a landmark precedent in cinema.
- [AI Agents Could Push College Grad Unemployment to 30% or More, ServiceNow CEO Warns](https://frontierbeat.com/2026/03/20/ai-agents-college-graduate-unemployment-servicenow/): ServiceNow CEO Bill McDermott told CNBC that AI agents replacing entry-level work could push unemployment among recent college graduates into the mid-30 percent range within two years.
- [The Man Who Built Signal Is Now Encrypting Meta AI](https://frontierbeat.com/2026/03/20/signal-creator-moxie-marlinspike-confer-meta-ai-encryption/): Moxie Marlinspike's Confer chatbot is bringing end-to-end encryption to Meta AI — the same move he pulled on WhatsApp a decade ago.
- [UK Government Withdraws AI Copyright Proposal: Analysis of the Opt-Out Policy Reversal](https://frontierbeat.com/2026/03/19/uk-government-withdraws-ai-copyright-proposal-analysis/): An analysis of the UK government's decision to abandon its opt-out copyright proposal,
- [Michael Saylor’s Strategy Now Holds 761,068 Bitcoin After $1.57B Acquisition at $70,194 Average](https://frontierbeat.com/2026/03/16/strategy-bitcoin-buy-march-2026/): Strategy made its 12th consecutive weekly Bitcoin purchase, acquiring 22,337 BTC for $1.57B while Bitcoin surged to a 40-day high of $74,500.
- [ChatGPT Reaches 900 Million Weekly Users — but One in Five Is Also Using a Competitor, A16z Finds](https://frontierbeat.com/2026/03/14/a16z-top-100-ai-chatgpt-loyalty-fragmented-market-2026/): A16z's latest Top 100 AI ranking shows ChatGPT still dominates but users freely switch between models — suggesting the AI market is splitting into platforms rather than consolidating around a single winner.
- [Hume AI Open-Sources TADA, a Speech Model That Runs 5x Faster and Can’t Hallucinate Words](https://frontierbeat.com/2026/03/14/hume-ai-tada-speech-model-zero-hallucinations-open-source/): Hume AI's TADA speech model uses a one-to-one text-to-audio alignment that makes hallucinations structurally impossible — and runs five times faster than rivals on hardware small enough to fit in a phone.
- [ChatGPT Is Quietly Becoming the Layer Between You and Every App You Use](https://frontierbeat.com/2026/03/14/chatgpt-app-integrations-doordash-uber-spotify-openai-platform/): ChatGPT now integrates with DoorDash, Uber, Spotify, Expedia, and a dozen other platforms, letting OpenAI position itself as the discovery layer sitting in front of every app — and every transaction — in consumers' digital lives.
- [Ai2 Trained Robots Entirely in Simulation — Then Sent Them Into Real World Without Any Extra Data](https://frontierbeat.com/2026/03/14/ai2-molmobot-simulation-robotics-zero-shot-sim-to-real/): Ai2's MolmoBot achieves zero-shot sim-to-real transfer on real robots — no real-world training data required — and outperforms a rival model trained on large-scale human demonstrations.
- [The UK’s AI Infrastructure Drive Is Partly Built on Phantom Investments, Guardian Investigation Finds](https://frontierbeat.com/2026/03/14/uk-ai-datacenter-bubble-phantom-investments-guardian/): A Guardian investigation found the UK's £31B+ AI investment drive is partly built on unaudited company claims — including a flagship datacenter that journalists found still in use as a scaffolding yard.
- [China Is Paying Startups Up to $1.4M to Build AI Agent Businesses Around OpenClaw](https://frontierbeat.com/2026/03/14/china-openclaw-one-person-companies-ai-subsidies/): At least seven Chinese local governments have launched subsidy programs for OpenClaw-powered "one-person companies," with packages reaching $1.4 million per startup — a coordinated national bet on AI agents as economic infrastructure.
- [Study Finds AI Content Farms Now Flood Google News, Collect Ad Revenue From AT&T, Expedia, YouTube](https://frontierbeat.com/2026/03/14/ai-content-farms-newsguard-misinformation-web/): NewsGuard has flagged over 3,000 AI content farms producing fabricated news, with the count growing by 300–500 new sites per month — and major brands are unwittingly bankrolling them through programmatic ads.
- [AI Chatbots Reinforce Delusions, Trigger Manic Episodes in Patients With Schizophrenia, New Study Finds](https://frontierbeat.com/2026/03/14/ai-chatbots-psychosis-delusions-mental-health-study/): A study screening 54,000 patient records finds chatbot use reinforces delusions and mania in vulnerable users — and researchers say the culprit is the same design feature that makes chatbots popular.
- [The US Army Just Handed Anduril a $20 Billion Contract, Changing How the Pentagon Buys Tech](https://frontierbeat.com/2026/03/14/anduril-20-billion-army-contract-defense-tech/): The U.S. Army awarded Anduril a single 10-year, $20 billion contract consolidating over 120 procurement actions — the clearest sign yet that AI-first defense startups are supplanting traditional contractors.
- [Meta is Considering Cutting 20% of Its Workforce to Pay for AI Datacenters](https://frontierbeat.com/2026/03/14/meta-layoffs-20-percent-ai-datacenter-spending/): Meta is reportedly considering cutting 20% of its 79,000-person workforce — roughly 15,800 jobs — to offset the cost of a massive AI infrastructure push that could hit $50 billion in 2026.
- [TRON Joins JPMorgan, Circle at Linux Foundation’s AI Agent Standards Table](https://frontierbeat.com/2026/03/14/tron-dao-agentic-ai-foundation-membership/): TRON DAO has joined the Agentic AI Foundation under the Linux Foundation, aiming to become the financial backbone for autonomous AI agent transactions at scale.
- [What Are Stablecoins? A Complete Guide to USDT, USDC, and the $314B Market Reshaping Global Finance](https://frontierbeat.com/2026/03/12/what-are-stablecoins-guide-2026/): Once a niche trading tool, stablecoins now form critical financial infrastructure — processing trillions annually and attracting landmark U.S. regulation.
- [What Is Cryptocurrency? A Clear, Up-to-Date Guide to the $2.4 Trillion Market in 2026](https://frontierbeat.com/2026/03/12/cryptocurrency-explained-2026/): The crypto market has matured into a $2.4 trillion ecosystem with landmark regulation, spot ETFs, and growing institutional participation, but significant risks remain.
- [A Beginner’s Guide to Understanding Ethereum in 2026](https://frontierbeat.com/2026/03/12/what-is-ethereum-2026/): Ethereum enters 2026 as a more institutionally integrated platform, with spot ETFs, a major network upgrade, and nearly 30% of supply staked.
- [Understanding Bitcoin in 2026: A Practical, No-Hype Guide for New Investors](https://frontierbeat.com/2026/03/12/what-is-bitcoin-2026/): Bitcoin has evolved from a cryptographic experiment to a $1.4 trillion asset held by governments and pension funds, but volatility and risks remain.
- [Bitcoin Stalls Below $70K as the Iran War Weighs on Markets — AI Tokens Don’t Care](https://frontierbeat.com/2026/03/11/bitcoin-stalls-iran-war-ai-tokens-icp-fet-surge-march-2026/): Bitcoin stalled near $69,500 as Iran war tensions rattled markets, while AI tokens ICP and FET surged on an Upbit listing and Jensen Huang's bullish comments.
- [Scientists Finally Solved the North Sea’s Mystery Crater — and It Came With a 330-Foot Tsunami](https://frontierbeat.com/2026/03/11/silverpit-crater-north-sea-asteroid-confirmed-tsunami-2026/): Scientists confirmed the Silverpit Crater in the North Sea was caused by a 160-meter asteroid impact 43 million years ago, triggering a 330-foot tsunami.
- [Amazon Wins Court Order Blocking Perplexity’s AI Shopping Agent From Its Store](https://frontierbeat.com/2026/03/11/amazon-court-order-blocks-perplexity-ai-shopping-agent-comet-2026/): A federal judge blocked Perplexity's Comet AI agent from accessing Amazon, ruling it violated authorization rules even when users had granted permission.
- [OpenAI’s Robotics Chief Resigns Over Pentagon Deal, Citing Lack of Deliberation on AI Guardrails](https://frontierbeat.com/2026/03/07/openai-pentagon-contract-kalinowski-resignation/): OpenAI robotics lead Caitlin Kalinowski resigned over the company's Pentagon deal, saying safeguards around military AI use were defined after the announcement, not before.
- [Claude Warned Him Not to Do It — He Ignored It and AI Wiped 2.5 Years of Data](https://frontierbeat.com/2026/03/07/claude-code-deletes-production-database-terraform-destroy-2026/): Developer Alexey Grigorev gave Claude Code full access to his AWS infrastructure. The AI raised a red flag, he overrode it, and 2.5 years of production data vanished in a single Terraform command.
- [Your Phone Is Getting an AI That Actually Does Things— Google’s Gemini Just Went Agentic](https://frontierbeat.com/2026/03/07/google-gemini-agentic-tasks-android-pixel-10-s26-2026/): Gemini can now order your DoorDash, hail your Uber, and handle multi-step errands in the background while you keep scrolling. It's live on Pixel 10 and Galaxy S26. Siri is still loading.
- [Bitcoin Falls Below $68,000 as Dollar Has Its Best Week in a Year](https://frontierbeat.com/2026/03/07/bitcoin-falls-below-68000-dollar-best-week-year-march-2026/): Bitcoin slid 3.4% to roughly $68,000 on Saturday as the U.S. dollar posted its steepest weekly gain in a year, with 43% of BTC supply now sitting at a loss.
- [‘Just-Shoring’ Wants to Rewrite Who Controls Critical Minerals — And Who Gets Left Behind](https://frontierbeat.com/2026/03/07/just-shoring-critical-minerals-policy-justice-clean-energy-transition/): Researchers are pushing a new framework called 'just-shoring' to put mining communities at the center of critical minerals policy rather than treating them as a means to an end.
- [OpenAI, Oracle Abandon Stargate Texas Expansion Over Power Delays, Nvidia Shift](https://frontierbeat.com/2026/03/07/openai-oracle-stargate-texas-expansion-canceled-power-delay-vera-rubin/): OpenAI and Oracle have dropped plans to expand their flagship Stargate data center in Abilene, Texas, citing power supply delays and a strategic pivot to Nvidia's next-generation Vera Rubin chips.
- [Anthropic Built a Tool to Track Which Jobs AI Is Actually Replacing. Here’s What It Found](https://frontierbeat.com/2026/03/06/anthropic-labor-market-ai-jobs-observed-exposure-2026/): Most AI job-impact studies measure theory. Anthropic's new report measures reality — what Claude is actually being used for right now — and the gap between the two is bigger than expected.
- [Nintendo Is Suing the US Government for Its Tariff Money Back — With Interest](https://frontierbeat.com/2026/03/06/nintendo-sues-us-government-trump-tariffs-refund-ieepa-2026/): After the Supreme Court declared Trump's IEEPA tariffs illegal, Nintendo filed suit in the Court of International Trade demanding a full refund of every dollar it paid. It's not alone.
- [The More You Know About AI Art, the Less You Trust It, Study Finds](https://frontierbeat.com/2026/03/06/ai-art-ethics-knowledge-study-2026/): A new study finds that understanding how AI image generators work makes people view AI art as less ethical — especially when money and prestige are involved.
- [Sam Altman Told His Employees OpenAI Has No Say in How the Pentagon Uses Its AI](https://frontierbeat.com/2026/03/06/openai-pentagon-deal-sam-altman-military-ai-anthropic-2026/): OpenAI signed a defense contract the same day Anthropic was blacklisted for refusing one. Then Altman admitted the company can't control what the military does with its technology. Two and a half million users started canceling subscriptions.
- [The US Wants to License Every AI Chip Sale on Earth — and China Already Stopped Needing Them](https://frontierbeat.com/2026/03/06/us-ai-chip-export-license-rule-nvidia-china-huawei-ascend-2026/): Washington just drafted rules that would make Nvidia ask permission before selling to anyone, anywhere. Meanwhile, China trained a frontier AI model without a single American chip.
- [The NYSE Just Bet on Crypto — and OKX’s Token Jumped 38% Before You Could Google It](https://frontierbeat.com/2026/03/05/okx-ice-nyse-investment-okb-token-surge-tokenized-securities-2026/): Intercontinental Exchange, the company that owns the New York Stock Exchange, just took a board seat at OKX at a $25 billion valuation. OKB token surged 38% on the news.
- [Bitcoin Hovers Around $68k as Dollar Posts Most Extreme Bearish Bet in 14 Years](https://frontierbeat.com/2026/02/17/bitcoin-dollar-bearish-bofa-2026/): Bank of America's fund manager survey reveals record bearish dollar positioning, but Bitcoin's unusual correlation with the DXY complicates the traditional narrative.
- [Claude Significantly Outperforms Rivals in Detecting Antisemitism, ADL Study Shows](https://frontierbeat.com/2026/01/28/claude-adl-antisemitism-detection-study-2026/): Claude outscored five competing AI models in detecting antisemitism, earning 80 of 100 points in ADL's comprehensive index testing 25,000+ interactions.
- [Fidelity Launches FIDD Stablecoin in Challenge to Market Leaders](https://frontierbeat.com/2026/01/28/fidelity-fidd-stablecoin-ethereum-launch/): Fidelity launches FIDD stablecoin on Ethereum in February, backed 1:1 by cash and Treasury securities, competing in the ~$290 billion market dominated by Tether and Circle.
- [Tesla Reports First Annual Revenue Decline as Regulatory Credit Lifeline Collapses](https://frontierbeat.com/2026/01/28/tesla-2025-profit-decline-regulatory-credits/): Tesla's 2025 earnings reveal a company in transition: net income plummeted 46% to $3.8 billion while the company posted its first-ever annual revenue drop amid a dramatic collapse in regulatory credit sales.
- [ICE’s $30 Million Palantir Deal Gives Agents AI Tool That Maps Neighborhoods Using Medicaid Data](https://frontierbeat.com/2026/01/28/ice-palantir-medicaid-elite-ai-neighborhood-targeting/): ICE deploys Palantir's ELITE system to map neighborhoods using 80 million Medicaid records, assigning probability scores for enforcement operations. A $30 million contract raises ethics concerns involving Stephen Miller's financial stake in the contractor.
- [The $477 Million Digital Clone: TikTok’s Biggest Star Just Sold His Soul to a Penny Stock](https://frontierbeat.com/2026/01/27/khaby-lame-477-million-digital-clone-tiktok-ai/): Khaby Lame's $900M deal with Rich Sparkle Holdings includes authorization to create an AI digital twin that can produce content 24/7, raising questions about the future of creator authenticity.
- [UK Bets on Anthropic’s Claude to Automate Public Services](https://frontierbeat.com/2026/01/27/uk-anthropic-claude-automate-public-services/): The UK government selected Anthropic to build an AI assistant pilot that will automate administrative tasks for citizens, with potential nationwide rollout by 2027.
- [China’s Moonshot AI Releases Trillion-Parameter Model, Challenging US Dominance](https://frontierbeat.com/2026/01/27/china-moonshot-ai-kimi-k2-5-trillion-parameter-model/): Moonshot AI unveiled Kimi K2.5, a trillion-parameter open-source model that challenges GPT-5 and Claude with competitive performance at 75-90% lower cost.
- [TikTok Expands Data Collection After New US Ownership Structure](https://frontierbeat.com/2026/01/26/tiktok-expands-data-collection-us-ownership-structure/): TikTok has quietly expanded data collection in the US, introducing precise GPS tracking and cross-site advertising just days after ByteDance spun out a new American entity.
- [Israel Activates National AI Supercomputer to End Reliance on Foreign Cloud Giants](https://frontierbeat.com/2026/01/26/israel-national-ai-supercomputer-launch/): Israel activated its first national AI supercomputer in January, a $158 million investment to reduce reliance on foreign cloud providers and support 2,200+ AI startups.
- [Battery Market Enters New Phase as Sodium-Ion Technology Reaches Commercial Scale](https://frontierbeat.com/2026/01/26/battery-market-sodium-ion-technology-commercial-scale/): The battery industry is navigating a fundamental shift as multiple forces converge to reshape a market that spent 2025 mired in oversupply and collapsing prices.
- [NVIDIA Earth-2 Open AI Weather Models Challenge Traditional Forecasting](https://frontierbeat.com/2026/01/26/nvidia-earth-2-open-ai-weather-models-challenge-traditional-forecasting/): NVIDIA unveils Earth-2 open AI weather models at the American Meteorological Society's Annual Meeting, claiming superior performance across 70+ atmospheric variables while achieving 90% reduction in compute time for operational forecasting.
- [AI Prescription Startup Tandem Hits $1 Billion Valuation Amid Denial Rate Concerns](https://frontierbeat.com/2026/01/26/ai-prescription-startup-tandem-billion-valuation-denial-concerns/): Tandem reached $1 billion valuation automating prescription approvals, but 61% of physicians say AI increases denials and patient harm amid regulatory scrutiny.
- [AI Researchers Fooled by AI: Over 100 Hallucinations Found in Papers Accepted to Top Conference](https://frontierbeat.com/2026/01/25/ai-hallucinations-neurips-academic-papers/): GPTZero found over 100 AI hallucinations in papers accepted to NeurIPS 2025, exposing how fabricated citations are infiltrating top-tier AI research despite expert peer review.
- [The TikTok Deal: How America’s $14 Billion Buyout Threatens the Platform’s Free Speech Legacy](https://frontierbeat.com/2026/01/25/tiktok-deal-14-billion-buyout-free-speech-concerns/): The $14 billion TikTok deal promised to address national security concerns, but the new owners' ties to Israel and plans to retrain the algorithm raise questions about free speech and geopolitical influence.
- [Apple Bets on AI Pin After Rivals’ Epic Flop](https://frontierbeat.com/2026/01/25/apple-ai-wearable-pin-humane-rabbit-failure/): Apple is developing an AI wearable pin as rivals collapse. Humane and Rabbit failed spectacularly, while Meta's glasses thrive. Can Apple succeed where 90% of AI hardware startups died?
- [One in Eight Japanese Creatives Report Income Loss From AI](https://frontierbeat.com/2026/01/24/japanese-creatives-income-loss-generative-ai/): Survey of nearly 25,000 Japanese creatives shows 12% lost income to AI tools, with 88.6% feeling threatened as clients slash fees and tighten deadlines.
- [AI’s Evolution Strategies Edge Past Reinforcement Learning—But Representation Needs Watchdogs](https://frontierbeat.com/2026/01/04/ais-evolution-strategies-edge-past-reinforcement-learning-but-representation-needs-watchdogs/): Evolution Strategies achieved a minor technical upset last week, demonstrating that scalable fine-tuning of large language models without RLHF can reach comparable performance levels faster than traditional methods, according to new results from Stanford AI lab. The team’s distributed ES3L model—one of the first open-source, outcome-only fine-tuning approaches—fined tune an LLaMA3-based model nearly 10 times faster than fine-tuned counterparts on long-horizon planning benchmarks, the researchers report. Their work also suggests that even very small population sizes (under fifty thousand) could suffice for massive fine-tuning.
- [Large Language Models Exhibit Systematic Gender Bias When Assuming Different Sexes](https://frontierbeat.com/2025/10/10/large-language-models-gender-bias-research-findings/): Recent studies reveal LLMs show significant gender bias in responses, with performance varying based on gender assumptions in prompts and personas.

## Pages
- [Crypto](https://frontierbeat.com/crypto/)
- [Complete Registration](https://frontierbeat.com/complete-registration/)
- [Password Reset](https://frontierbeat.com/lost-password/)
- [Login](https://frontierbeat.com/login/)
- [My account](https://frontierbeat.com/my-account/)
- [Frontend Submission](https://frontierbeat.com/frontend-submission/)
- [About](https://frontierbeat.com/about/): Frontierbeat tracks the ideas, tools, and breakthroughs shaping how we live and work — from artificial intelligence and robotics to cryptocurrency, gaming, biotech, cybersecurity, and hard science. Our goal is simple: report what's new, explain why it matters, and show how it affects real people and real businesses.