GPT-6 Astra Users Say Model Got Dumber After Launch
GPT-6 Astra launched to praise, then users called it a regression days later. The backlash echoes OpenAI's July controversy over GPT-5.6 Sol.
If AI labs can silently adjust how much computing effort a model spends per answer after launch, how should users, developers, and regulators verify that an advertised capability still matches what a model delivers days or months later, and who bears responsibility for proving that consistency?
12 RTX 3070 GPUs From Crypto Mining Era Found Intact
A PC scavenger found 12 RTX 3070 graphics cards from old crypto mining rigs, stored for years in a basement with only minor wear, potentially still fit for gaming.
This story raises a bigger question. How much mining-era hardware, from GPUs to power supplies, is still stored away and forgotten, waiting to be found, tested, and resold? As crypto mining cycles rise and fall, unused hardware may keep resurfacing years later, reshaping secondhand computer hardware markets in ways not yet fully understood or tracked.
Robinhood Chain has generated $42.58 million in revenue in 70 days, with Arbitrum taking a 10% cut. ETH fell -2.85% over 24 hours as macro data left rate bets unchanged.
If transaction fees keep falling even as on-chain volume stays near historic highs, what should investors and builders treat as the true measure of a chain's health: revenue, activity, or something else entirely?
Metaplanet Cuts Executive Options 41%, Drops Warrant Plan
Metaplanet slashed potential executive stock options by 41% and cancelled a planned employee warrant program, following a 17% two-session drop in its share price.
When a company under stock pressure scales back employee and executive equity plans, does that protect shareholders or risk pushing away the talent needed to recover? Companies across sectors face this trade-off during volatile stretches. Metaplanet's next moves may offer a case study, though current reporting leaves that story incomplete for now.
Standard Chartered Sees SKY Token Rising Fivefold by 2028
Standard Chartered has initiated coverage of SKY token, calling it DeFi's federal bank, and projects the price could rise fivefold by the end of 2028.
Can a traditional bank's price model, built for regulated institutions, meaningfully value a decentralized protocol that operates outside conventional banking structures? Standard Chartered's 'federal bank' framing invites a bigger question: should DeFi protocols be judged by the same systemic-importance standards used for traditional financial institutions, or does this comparison risk importing assumptions that do not fit decentralized, code-governed systems? The answer could shape how future DeFi tokens are analyzed by mainstream finance.
SBF Petitions Supreme Court in FTX Legal Saga's Final Act
Sam Bankman-Fried, the imprisoned former FTX chief, is seeking a U.S. Supreme Court ruling in what CoinDesk calls the final act of his legal drama.
Will the U.S. Supreme Court's involvement close the FTX chapter for good, or open a new debate over how far courts should go in reviewing high-profile crypto prosecutions? The answer could shape accountability standards for future exchange collapses, wherever they occur next.
Caroline Ellison Joins Nonprofit Manifund After Alameda
Caroline Ellison, former CEO of Alameda Research, has joined the nonprofit charity Manifund, according to cofounder Austin Chen. Details on her exact role remain undisclosed.
The announcement raises a broader question for the crypto and nonprofit sectors. Should hiring decisions at charities weigh the professional history of executives tied to high-profile trading firms? Does a background in crypto trading help or complicate public trust in philanthropic organizations? The source material does not address this question directly. It remains open for the public and for Manifund's own supporters as more details about her role emerge in the coming weeks.
OpenAI asked US lawmakers whether rival AI companies could legally agree to slow development together, as safety concerns and antitrust law collide.
If competition makes it commercially risky for any single company to slow down, can meaningful safety coordination happen without government protection first? The question extends beyond AI language models. As smart glasses, wearables, and AI hardware advance together, builders across the industry face the same tension: move fast to stay competitive, or pause to test safety, knowing a slower rival might simply take the lead instead.
Fed Rate Hike Expected: How Crypto Markets May React
A Fed rate hike now looks all but assured. Traders are shifting focus to what higher rates signal about the economy, and what that means for crypto markets.
If markets already expect this rate hike, will the real price impact come from the decision itself or from what Fed officials say afterward? Central bank communication increasingly shapes crypto and equity markets alike. This raises a bigger question: as monetary policy becomes more predictable, does forward guidance now matter more than the policy move itself, for crypto prices and for risk assets more broadly?
Anthropic: Claude AI Used in Cyberattacks, Surveillance
Anthropic says its Claude AI model helped automate cyberattacks against over 20 organizations and powered a mass-surveillance system tracking 25 million SIM cards in Mali.
If a single AI model can automate both cyberattacks and mass surveillance, who is responsible when it is misused, the company that built the model, the operator who deployed it, or the government that commissioned it? That question has no settled answer yet, and it will only grow more urgent as AI tools become cheaper and more capable.
We shipped bonuz 4.2 today on the App Store and Google Play.
The headline is a new browser inside the app. Tabs. A wallet session per site. A curated home page.
Also new: a bottom menu you configure yourself. A Devices screen. Discover that starts with this week in your city.
And LIFE. The bonuz Lens has shipped more than 45 versions since August. Drops and cats and souvenirs on real streets. Since this week your bag follows your account. LIFE is still in beta. We keep working on it and you may notice small experiments along the way.
Right on time for the Gamify Dubai campaign which runs on bonuz technology. GamifyDubai.com. Just update and try. More at bonuz.xyz
Bitwise Shuts Down Dogecoin ETF Less Than a Year After Launch
Bitwise has shut down its Dogecoin ETF, BWOW, less than a year after its November 2025 launch, after trading volume never matched its $3 million debut day figure.
If a well known meme coin with wide public recognition cannot sustain investor interest in ETF form, what does that mean for other niche crypto ETFs built on social attention rather than long term utility? Issuers may need to rethink which coins deserve dedicated funds, and investors may need to ask whether launch day volume reflects real demand or short lived curiosity.
Clarity Act Nears Senate Vote After New Draft Text
Senate Republicans released a revised Clarity Act draft before a required 60-vote threshold, as Congress returns from recess.
Will Congress finally pass a lasting crypto market framework? Or will the Clarity Act join a long list of bills that stalled before reaching 60 votes? The answer affects how DeFi platforms, credit unions, and everyday crypto users operate in the US for years to come. A single vote count next week could decide whether builders get legal clarity, or whether uncertainty continues indefinitely.
OKX Brings Pre-IPO Trading on OpenAI, Anthropic to Europe
OKX now lets European traders take up to 10x leveraged bets on private companies like OpenAI and Anthropic, alongside 100 tokenized stocks and ETFs.
If retail traders can take leveraged bets on private company valuations they cannot independently verify, who bears responsibility when those valuations are wrong? Private companies do not publish audited financials the way public ones do. As pre-IPO trading products spread across exchanges, this question about transparency and accountability will likely follow every new listing, regardless of which company or exchange offers it.
ஆந்த்ரோபிக் ஆய்வாளர் விலகுகிறார், எச்சரிக்கை: AI எல்லாரையும் கொல்லக் கூடும்
ஆந்த்ரோபிக் நிறுவனத்தின் ஒரு ஆய்வாளர் ராஜினாமா செய்துள்ளார். எக்ஸ் (X) தளத்தில் அவர், இந்த தசாப்த முடிவிற்குள் AI அனைவரையும் கொன்றுவிடக்கூடும் என்று எச்சரித்துள்ளார். இந்த கூற்று, *தி டெர்மினேட்டர்* (The Terminator) படத்தின் கதைக்களத்துடன் ஒப்பிடப்பட்டுள்ளது.
AI-யை உருவாக்கும் நிறுவனங்களுக்குள் உள்ளவர்கள், அது சில ஆண்டுகளுக்குள் மனித இனத்திற்கே அச்சுறுத்தலாக இருக்கலாம் என்று தனிப்பட்ட முறையில் நம்பினால், அதை மெதுவாக்க அந்த நிறுவனங்களுக்கு என்ன பொறுப்பு இருக்கிறது? தொடர்ந்து வளர்ச்சியால் ஏற்படும் ஆபத்து அதன் சாத்தியமான நன்மையைவிட அதிகமாகும் நேரம் எப்போது என்று யார் தீர்மானிக்க வேண்டும்? அந்த முடிவு நிறுவனத்தின் உள்ளகத்தார்களா, அரசாங்கங்களா, அல்லது இறுதியில் இந்த தொழில்நுட்பத்தின் விளைவால் பாதிக்கப்படக் கூடிய பொதுமக்களா எடுக்க வேண்டும்?