Ontology said it received an urgent report that a project built on the Ontology blockchain carries a potential security risk that could threaten assets on the network. As a precaution, it coordinated with the majority of consensus nodes, and the nodes agreed to temporarily stop block production to support the investigation and reduce potential risk.

Halting block production on a live layer-1 means on-chain transactions cannot be processed for the duration of the review. Ontology frames the halt as a precaution pending the investigation, and no confirmed exploit or loss figure has been reported.

Ontology said the network will resume block production once the investigation makes sufficient progress and safety is confirmed, and that it will keep working with relevant parties and closely monitor the situation. The project involved has not been publicly identified and no loss figure has been given.

Aerodrome and Velodrome will merge into a single cross-chain DEX called Aero, now scheduled to launch on October 21 across Base, Ethereum mainnet, OP, Arc and Ink, with Robinhood Chain and Arbitrum newly added. AERO and VELO will be consolidated into a single AERO token, with roughly 94.5% of prior revenue distributed to AERO holders and 5.5% to VELO holders; on October 11 HTX quotes showed AERO at $0.9422 (+17.29% over 24 hours) and VELODROME at $0.04087 (+16.74%).

The merger folds two widely used DeFi governance and liquidity tokens into one AERO token and combines their liquidity across multiple chains; according to AMBCrypto, Aero aims to capture 10–15% of combined Layer 2 DEX volume, equivalent to more than $2 billion in monthly trading activity.

The revenue split is based on prior income, at roughly 94.5% to AERO holders and 5.5% to VELO holders, and the community and trading platforms are also migrating to Aero.

Nvidia is in early-stage talks to acquire US open-weights model developer Reflection AI or to increase its investment in the company, people familiar with the matter said, according to a Financial Times report cited by BlockBeats on October 11. The deal could take several forms, including an acqui-hire, further equity investment, or agreements to supply Reflection with more chips and compute, and the two companies could reach an agreement within weeks.

Nvidia is already one of Reflection's largest shareholders, and an acqui-hire structure would let it absorb the startup's staff and license its technology without the lengthy regulatory review a full acquisition typically requires.

The talks are still at an early stage and no agreement has been reached. Reflection AI was valued at $25 billion in a funding round in March.

Last week a range of parties — including former Senator Chris Dodd, former CFTC and SEC Chair Gary Gensler, the NFL, 39 states plus the District of Columbia and 145 tribal nations — filed amicus briefs urging the U.S. Supreme Court to take up the question of whether sports-based prediction market contracts are swaps. On Friday the CFTC issued two proposals on the same question, one of which is an interim final rule that is already in effect.

There is now a major circuit court split: two appellate courts have ruled that prediction markets touching sports are gambling products that should be regulated by the states, while one appellate court ruled the matter falls to the CFTC.

Dodd's brief argues Dodd-Frank was not intended to give the CFTC authority to usurp state gambling regulation, and that parlays chaining multiple wagers do not facilitate hedging or price discovery and therefore are not swaps under the Commodity Exchange Act; he also disputed that the CFTC's grant of "exclusive jurisdiction" is an express preemption provision. According to the CFTC, the interim final rule codifies its longstanding position that casino-style gambling products, including sportsbook and casino wagers, are excluded from the "swap" definition, while a separate proposal would further define swaps to include event contracts.

A trader netted $1.28 million by pushing Hyperliquid's ETH prices through Papertrade, a newly launched exchange offering up to 1,000x leverage, according to the report. The same wallet later lost $450,000 when 30 short positions were liquidated on Saturday night after having been held open for more than three hours.

The report frames the episode as a concrete cross-venue price-manipulation vector touching Hyperliquid, one of the largest perpetual futures DEXs, rather than a routine trading loss.

The 30 short positions were liquidated on Saturday night, more than three hours after they were opened.

The IMF's latest Global Financial Stability Report found that more than 50% of tokenized stock trading occurs outside regular U.S. market hours and that about 80% of trades are for less than a full share, while tokenized equities were roughly 1.5 times more volatile and less liquid than the underlying stocks. The report, covering the five most liquid tokenized U.S. equities across 11 trading venues over 365 days, says systemic risks remain limited for now and urges regulators to consider circuit breakers for 24/7 trading.

The IMF said it urged regulators to consider circuit breakers for 24/7 trading and to monitor links between tokenized and traditional markets more closely, while noting that systemic risks remain limited for now. It also said its findings "should be interpreted with caution" given the market's early stage.

The tokenized equity market is worth about $2.3 billion and highly concentrated, with Ondo Finance and Backed Finance's xStocks accounting for more than 70% of it; both issue synthetic tokens offering price exposure without direct share ownership. When traditional markets reopened, conventional shares absorbed 87% to 99% of the overnight price moves their tokenized counterparts had made, and tokenized prices on decentralized exchanges strayed further from traditional markets.

Developer momo5502 published a blog post describing an experiment in which AI agents consumed roughly 500 billion tokens to decompile a commercial first-person shooter down to byte-matching assembly. The post, which drew a large Hacker News thread, discusses the cost, the agent architecture used, and whether functional equivalence would have been enough.

The post provides a concrete, large-scale datapoint on the cost of agentic LLM workflows, attaching a 500-billion-token figure to a single reverse-engineering project. Commenters read it as evidence that byte-matching decompilation is a deliberately expensive target compared with functional equivalence.

Byte-matching (matching) decompilation means writing original C++ that, compiled with the era-correct toolchain, produces byte-for-byte the same machine code as the shipped binary. According to a commenter quoting the post, two earlier posts by the author have since been removed.

Terence Tao has published the slides for his Caltech 2026 talk "Math 2.0", which lays out the historical "Math 1.0" state of mathematics and the current state that AI is inducing in the field. Tao's blog post accompanying the talk is dated 10 October 2026 and notes the Caltech talk was postponed to October and changed significantly in light of recent events.

The slides offer a prominent mathematician's framing of how AI is changing mathematical practice and theory, and the accompanying Hacker News thread drew unusually heavy engagement (288 points, 349 comments).

The slides reference the Inverse Galois Problem challenge (2026), whose first stage, IGP24, was a crowdsourced competition to locate degree-24 polynomials covering each of the 25,000 possible Galois groups at that degree. Tao's blog post is filed under math.GM as a talk entry.

A Simply Wall Street analysis reports that a CFTC rule change is poised to affect Coinbase and two other trading platforms, drawing regulatory and stock-market attention. The piece is framed as stock analysis; the other two platforms are not named in the headline or summary.

Per the CFTC's own description, the proposed Regulation CTX and Regulation CAM would establish requirements for CFTC-registered exchanges that offer crypto assets for trading, which is the compliance channel through which listed US crypto venues could be affected.

The proposed rulemaking package, titled "Regulation Crypto Asset Transactions and Regulation Crypto Asset Markets," was submitted to the White House Office of Information and Regulatory Affairs on September 17, 2026 for regulatory review. In a CFTC-published statement, the agency said these regulations would not require crypto assets to trade on CFTC-registered platforms, a step it said it lacks authority to impose without congressional action.

Nasdaq-listed SVRN announced it has acquired FastNEAR, a provider of NEAR RPC, archival and data infrastructure; the purchase price was not disclosed. FastNEAR becomes a wholly owned subsidiary and SVRN's first operating business alongside its asset treasury, with FastNEAR co-founders Evgeny Kuzyakov and Mike Purvis joining SVRN.

The deal moves a public-market company beyond treasury-holding into operating crypto infrastructure, and places stewardship of widely used NEAR endpoints and data services under a Nasdaq-listed owner.

FastNEAR handles more than 3.5 billion requests per month and provides over 100 TB of data services. After the acquisition, its RPC service, archival infrastructure and NEARDATA will continue to be operated by the original team, and SVRN said it plans to invest further in system reliability and service capacity.