TechFlow Intelligence Bureau: ChatGPT Helps Amateur Mathematician Crack 60-Year-Old Problem, CFTC Sues New York Regulator Over Coinbase and Gemini

marsbitОпубліковано о 2026-04-26Востаннє оновлено о 2026-04-26

Анотація

An amateur mathematician, with the assistance of ChatGPT, has solved a combinatorial mathematics puzzle originally proposed by Hungarian mathematician Paul Erdős in the 1960s. This marks another milestone in AI-aided mathematical research, demonstrating the evolving capabilities of large language models in formal reasoning. In other AI developments, OpenAI introduced a new privacy filter tool for enterprise API usage, automatically screening sensitive data. Meanwhile, the Qwen3.6-27B model achieved 100 tokens per second on a single RTX 5090 GPU using quantization, significantly lowering the cost barrier for local AI deployment. In crypto and Web3, the U.S. CFTC sued New York’s financial regulator, challenging its oversight of Coinbase and Gemini—a first-of-its-kind federal-state regulatory clash. Following a vulnerability, KelpDAO and major DeFi protocols established a recovery fund. Tether froze $344 million in assets linked to Iran’s central bank upon U.S. Treasury request, highlighting the centralized control risks in stablecoins. Separately, Litecoin underwent a 3-hour chain reorganization to undo a privacy-layer exploit. In the U.S., former President Trump invoked the Defense Production Act to address power grid bottlenecks affecting AI data centers and dismissed the entire National Science Board, raising concerns over research independence. A retail trader gained 250% on a $600k Intel options bet amid AI-related speculation. Xiaomi announced its first performance el...

An amateur mathematician used ChatGPT to solve a combinatorial mathematics problem proposed by Hungarian mathematician Erdős in the 1960s. This marks another milestone in AI-assisted mathematical research, demonstrating that large language models' capabilities in formal reasoning are evolving from "auxiliary verification" to "substantive breakthroughs."

Hot debate: The HN community is arguing whether this counts as a victory for "human+AI" or a pure AI breakthrough, as well as the issue of the rigor of the mathematical proof.

Scientific American | HN Discussion

OpenAI Launches Privacy Filter

OpenAI released a new privacy filtering tool that allows enterprise users to automatically filter sensitive information when calling the API. This is another step in OpenAI's compliance efforts for the enterprise market, following its data retention policy.

OpenAI Official

Qwen3.6-27B Achieves 100 tps on a Single RTX 5090

Using vllm 0.19 and INT4 quantization, Qwen3.6-27B achieved 100 tokens/second with a 256k context length. Consumer-grade GPUs are approaching the performance of last year's enterprise-grade inference cards, lowering the cost barrier for local deployment yet again.

Reddit Discussion

> Spicy take: Last year we were discussing "cloud vs. local," this year it's "which consumer card offers the best bang for the buck." The speed of AI compute moving downstream is faster than anyone expected.

Crypto / Web3

CFTC Sues New York Regulator Over Coinbase and Gemini

The U.S. Commodity Futures Trading Commission has filed a lawsuit against the New York State Department of Financial Services, alleging its regulation of Coinbase and Gemini interferes with federal authority. This is the first direct conflict between U.S. federal and state regulators in the crypto space and could reshape the U.S. crypto regulatory landscape.

Finance Feeds

DeFi Community Establishes Recovery Fund After KelpDAO Exploit

Following an exploit on KelpDAO that caused rsETH to depeg, protocols including Aave, EtherFi, Ethena, and Lido jointly established a recovery fund, with KelpDAO contributing 2000 ETH. A final plan has been determined, pending a DAO vote and technical execution; 71,700+ users affected.

Hot debate: The community is discussing whether this is the "DeFi mutual aid spirit" or "large protocols forced to save themselves to prevent systemic risk."

Aave Twitter | KelpDAO Twitter

Tether Freezes $344 Million in Iranian Central Bank Wallets

The U.S. Treasury's OFAC blacklisted two crypto wallets directly associated with the Iranian central bank, and Tether complied by freezing $344 million. This is the largest single amount frozen by a stablecoin issuer enforcing sanctions, again exposing the "kill switch" risk of centralized stablecoins.

Crowdfund Insider

Brazil Bans Kalshi and Polymarket

Brazilian regulators banned prediction market platforms like Kalshi and Polymarket, citing the need to "protect investors from the risks of unlicensed gambling." Following China and India, another emerging market says no to decentralized prediction markets.

Crowdfund Insider

Litecoin Rewrites 3 Hours of History to Undo Privacy Layer Exploit

The Litecoin network experienced its first major privacy layer exploit leading to double-spending; miners coordinated to reorganize the chain and roll back 3 hours of transaction records. This is another collision between the "immutable" promise and real-world security.

The Block

> Spicy take: Tether freezing $344 million just takes a backend operation; Litecoin rolling back 3 hours requires coordination from the entire miner network. Sometimes, "decentralization" is actually more fragile.

U.S. Stocks

Trump Invokes Defense Production Act to Address Grid Equipment Bottlenecks

Trump used the Defense Production Act to accelerate grid equipment production and energy project approvals, aiming to alleviate power supply bottlenecks for data centers and AI compute. This is the first time the U.S. government has elevated AI infrastructure to a "defense-level" priority.

Reddit Discussion

$600k Intel YOLO Position Rises 250%

A WSB user shared a screenshot showing a $600,000 Intel position with a 250% gain. Intel's stock price has recently rebounded on rumors about its AI chips and foundry business, but its fundamentals haven't fundamentally improved yet.

Reddit Discussion

Tech Companies

Lei Jun Announces Xiaomi YU7 GT Launch for Late May

The performance version of the Xiaomi car, the YU7 GT, will be launched in late May, positioning itself against the Tesla Model Y Performance and the Li Auto L6. This is Xiaomi's first performance model since entering the new energy vehicle赛道 (track/arena).

Zhihu Discussion

Trump Fires Entire National Science Board

The Trump administration fired all members of the U.S. National Science Board (NSB). The NSB provides policy advice to the National Science Foundation, and this move has raised concerns in academia about funding for basic research and its independence.

The Verge

> Spicy take: Silicon Valley is scrambling for H100s while watching the White House clear out its science advisory board. The AI boom and a basic research winter are happening simultaneously in the U.S. in 2026.

New Products / New Trends

iPhone Automatically Installs Mysterious App Daily

An HN user reported that their iPhone silently installs the same app every day, and it reappears no matter how many times it's deleted. Apple has not yet responded; suspicion points to MDM (Mobile Device Management) or an enterprise configuration profile.

HN Discussion

Finance / Macro

A-Shares See Second-Longest "Grouping" Since 2007

A CITIC Securities research report points out that the current duration of institutional "grouping" in A-shares is the second longest since 2007, suggesting the next allocation direction should be "neither hot nor contrarian." Core asset valuations are already at historical highs, and signals for a style rotation are strengthening.

Wall Street News

Today's Undercurrent

While ChatGPT helps an amateur mathematician crack an Erdős problem, Trump fires the entire National Science Board; Tether freezing $344 million in Iranian wallets proves the stablecoin "kill switch" is real, yet Litecoin relies on full-network coordination to roll back 3 hours to fix a vulnerability. The tech world in 2026 is playing out two parallel narratives: AI gives individuals unprecedented capabilities, but macro power structures (whether governmental or centralized protocols) are becoming more explicit and direct. "Decentralization" is not a technical problem; it's a political economy problem.

Пов'язані питання

QWhat major mathematical problem did an amateur mathematician solve with the help of ChatGPT, and who originally proposed it?

AAn amateur mathematician used ChatGPT to solve a combinatorial mathematics problem proposed by Hungarian mathematician Paul Erdős in the 1960s.

QWhich U.S. federal agency sued the New York State regulator over its oversight of Coinbase and Gemini, and why is this significant?

AThe U.S. Commodity Futures Trading Commission (CFTC) sued the New York State Department of Financial Services, marking the first direct conflict between federal and state regulators in the crypto space, potentially reshaping U.S. crypto regulation.

QWhat action did Tether take regarding Iranian central bank wallets, and what risk does this highlight?

ATether froze $344 million in wallets linked to the Iranian central bank after they were blacklisted by the U.S. Treasury OFAC, highlighting the 'switch' risk of centralized stablecoins.

QHow did the Litecoin network address a major privacy layer vulnerability, and what does this incident reveal about decentralization?

ALitecoin miners coordinated to reorganize the chain and roll back 3 hours of transaction history to fix a privacy layer bug that caused double-spending, showing that decentralization can sometimes be more fragile in practice.

QWhat tool did OpenAI release for enterprise users, and what is its purpose?

AOpenAI released a privacy filtering tool that allows enterprise users to automatically filter sensitive information when calling its API, as part of its compliance efforts in the enterprise market.

Пов'язані матеріали

From South Korea to the United States: Blue-Collar Jobs Are Becoming Increasingly Popular, Thanks to AI

AI is reshaping the labor market's value proposition. The traditional four-year college degree is losing its appeal as a guaranteed career path, while skilled blue-collar trades like electricians, welders, and plumbers are experiencing historic demand and wage premiums. This shift is driven by dual pressures: AI's displacement of certain white-collar roles and a booming need for physical infrastructure and data center construction. Data confirms the trend. In the U.S., vocational school revenue surged, and a significant portion of recent layoffs are AI-related. Surveys show a majority of Gen Z adults plan to pursue blue-collar work, citing better job security against AI automation. Vocational education interest has exploded recently. Experts cite a psychological shift as younger generations seek tangible, AI-resistant careers and avoid high student debt. In many cases, salaries for skilled trades now match or exceed those requiring a bachelor's degree. In South Korea, semiconductor vocational high schools boast near-total employment, with graduates securing high-paying roles at companies like Samsung. The shortage is structural, exacerbated by a retiring baby boomer workforce and massive infrastructure projects. Companies like JPMorgan Chase, Meta, and Lowe's are investing heavily in training programs. However, overcoming historical stigma and a "perception gap" around trade careers remains a key challenge to closing the talent gap.

marsbit47 хв тому

From South Korea to the United States: Blue-Collar Jobs Are Becoming Increasingly Popular, Thanks to AI

marsbit47 хв тому

Qualcomm: AI Hype Subsides, When Will Smartphones Emerge from the Gloom?

Qualcomm reported its Q3 FY2026 results (ending June 2026), with revenue of $9.95B, down 4% YoY but above expectations. Gross margin declined to 53.1%, pressured by rising costs across manufacturing and memory. Key business segments showed mixed performance: Handset revenue fell 19.6% YoY to $5.09B, dragged by an 11% decline in non-Apple Android shipments and weaker high-end mix. Conversely, Automotive revenue surged 61% to $1.59B, and IoT grew 9% to $1.83B. Core operating profit dropped 41% YoY due to margin compression and higher expenses. Management's Q4 FY2026 guidance projects revenue of $9.7B-$10.5B, in line with consensus, but Non-GAAP EPS guidance of $2.05-$2.25 fell short of expectations. Amidst persistent weakness in its core handset market, Qualcomm is pursuing growth in AI, focusing on Edge AI (smartphones, PCs, automotive) and Data Center AI. Its data center strategy includes four pillars: AI accelerators (e.g., AI200), commercial CPUs (Dragonfly C1000), custom silicon, and connectivity solutions. While these initiatives initially boosted its stock, concerns over AI capital expenditure sustainability have since erased those gains. The company targets $5B in data center revenue for FY2027 and $15B for FY2029. The report concludes that with the traditional handset business still under pressure, the data center opportunity is currently viewed as a longer-term option, and a more conservative valuation based on core operations may be warranted until AI contributions materialize.

marsbit51 хв тому

Qualcomm: AI Hype Subsides, When Will Smartphones Emerge from the Gloom?

marsbit51 хв тому

From TPU to Self-Evolving Agents: How Jeff Dean Predicts the Next Step in AI

At the 2026 YC Startup School, Jeff Dean outlined his vision for AI's next phase, shifting focus from simply scaling models to building intelligent, autonomous systems. He believes AI's progress is no longer just about creating smarter models, but about integrating them into systems capable of long-term, iterative work, automated experimentation, and continuous learning. This evolution moves the competition from "who has the bigger model" to "who can best organize intelligence." Dean suggests AI capabilities are now comparable to a junior engineer, enabling the automation of complex workflows. However, the true challenge and opportunity lie in managing these AI "workers" at scale. He emphasizes the importance of **context engineering**—structuring tools, memory, and feedback loops—over raw model power. For startups, this means building deep expertise in niche domains where general models currently fail (near 0-1% success rates), leveraging proprietary data, specialized tools, and domain-specific evaluators. A recurring theme is re-examining fundamental constraints. Dean's past work, like moving Google's search index to memory or creating the TPU, stemmed from questioning outdated assumptions about hardware and cost. He sees similar inflection points today, particularly in **specialized inference hardware** to drastically reduce latency and energy consumption for real-time Agent operation. Notably, he points out that in modern AI systems, the dominant cost is often not computation but **data movement**. Reliable, long-running Agents require robust system design, borrowing concepts from distributed computing like checkpointing, state management, and parallel exploration to handle failures and maintain progress over days or weeks. As AI automates execution, the scarcest human skills will shift to **defining clear specifications**, **judging what problems are worth solving** (taste), and designing effective feedback loops. Ultimately, Dean's framework prioritizes understanding the problem deeply, identifying the true bottlenecks, and systematically building closed-loop systems where AI can not only perform tasks but also improve AI itself.

marsbit51 хв тому

From TPU to Self-Evolving Agents: How Jeff Dean Predicts the Next Step in AI

marsbit51 хв тому

Торгівля

Спот
活动图片