Perplexity open-sources two decision models, with costs as low as 1 cent per million input tokens and inference speed boosted by up to 5 times.

Insight Beating AI Flash News, according to Politico, Anthropic is privately lobbying U.S. lawmakers on a draft cutting-edge AI security bill that sets strict thresholds for government intervention in AI models. Under the draft, the government may only seek to shut down an AI model if it faces foreseeable and imminent specific risks that could cause over $1 billion in losses, or lead to more than 50 deaths or serious injuries. The draft limits regulatory triggers to three risk categories: AI models used for bioweapon development or proliferation, models escaping developer control, and models triggering cyberattacks. The bill also requires AI developers to conduct transparent testing and undergo independent third-party assessments. If an assessment body identifies relevant risks, it must report to the U.S. Department of Commerce within 72 hours, and the Commerce Department may apply to a federal court to suspend the model’s operations, with the court making the final ruling. Additionally, the draft proposes to prohibit other U.S. government departments and officials from independently halting the development, deployment, or internal use of AI models. Anthropic stated that the proposal is derived from its advanced AI framework released earlier this year, aiming to require developers to undergo security assessments while empowering the government to block the deployment of dangerous models. The draft is currently in the lobbying phase and has not yet become law. If the relevant provisions are ultimately passed, they could significantly limit the U.S. government’s authority to take emergency interventions against cutting-edge AI models.

Crypto KOL CM noted that the market may have already priced in the impact of MSCI’s removal of MicroStrategy (Strategy), meaning MSTR’s stock price may not see further sharp declines when actual passive selling pressure emerges. MSTR closed down 9.56% yesterday. Current valuations of MSTR use different metrics. The standard mNAV is the ratio of the company’s market capitalization to its BTC holdings value, while net mNAV accounts for factors like preferred shares and debt, reflecting the stock price relative to the net BTC value per share. Net mNAV readings below 1 are rare, typically signaling the market’s growing concern over the company’s operational health or its valuation logic. Take June this year as an example: back then, the market worried about the company’s insolvency due to Strategy’s tight cash reserves and falling BTC prices, leading to downward pressure on net mNAV at the time. Currently, Strategy’s cash reserves are ample, theoretically enough to cover its dividend payments for the next three years, and BTC prices remain above its roughly $75,000 cost basis for BTC holdings. As such, the likelihood of net mNAV falling back below 1 is low. If it does drop below 1, that could instead present an investment opportunity.

According to Lookonchain monitoring, trader 0xC7Fd opened a 2x leveraged long position on 20.5 million AGT tokens on Aster DEX, with the position valued at approximately $1.63 million. The unrealized profit on this position has now exceeded $1.26 million, translating to a 155% return.

A Solana Foundation official stated that the U.S. GENIUS Act provides regulatory impetus for enterprises and financial institutions to enter the stablecoin sector, and is also one of the key background factors for Samsung to launch its stablecoin payment service. Samsung has announced that it will start supporting cross-border USDC transfers in Samsung Wallet during the last week of October, with the first phase targeting eligible Galaxy users in the U.S., covering around 82 million devices. Users can transfer dollar stablecoins to compatible crypto wallets in seconds with no transfer fees, and also remit to eligible bank accounts in over 60 countries and regions. The service will use Solana and Sui networks as underlying infrastructure, with transfer services provided by licensed stablecoin custodian Bastion, and Coinbase Prime acting as a secondary custodian responsible for safeguarding USDC assets. Samsung and Solana have been preparing for the relevant cooperation for many years, and the two sides will advance product implementation in phases, prioritizing increasing user adoption, and may expand to online and offline stablecoin payment scenarios in the future.

According to monitoring by AI Aunt, a whale address that opened a $10 million altcoin position starting September 23 withdrew 2.455 million ONDO from Binance 12 hours ago, valued at approximately $1.144 million with an average withdrawal price of $0.4662. The address currently holds around $9.218 million in three altcoins on-chain: 1.014 million UNI (worth ~$8.91 million, average withdrawal price $8.79, unrealized loss ~$1.551 million), 2.455 million ONDO (valued at ~$1.144 million), and 12,400 LTC (worth ~$841,000, average withdrawal price $67.86, unrealized loss ~$101,000). Additionally, the address’s BNB holdings were apparently liquidated six days ago, with an estimated loss of ~$73,000.

Beating AI News: Elorian AI, founded by former Google DeepMind researcher Andrew Dai, has launched its first visual reasoning model Glimpse, designed to handle actions, spatial relationships and object changes in videos. Glimpse can recheck frames during inference, zoom in on local areas, revisit key clips, and run code to verify details when needed. For example, to determine if a train is slowing down, it tracks the position of the same window across frames to compare displacement, rather than guessing based solely on video content. While similar frame cropping and zooming technologies have existed before, Elorian has made repeated checks of visual evidence a core capability of the model. Elorian’s internal tests show Glimpse outperforms GPT-6 Astra in four benchmarks: scientific chart reasoning, short video description, object counting and robot operation event recognition; it ties Astra in one test and lags in six others. Notably, Glimpse scores 56.68 on the comprehensive video understanding benchmark Video-MME v2, lower than Astra’s 64.06. These results are self-reported by the company and have not yet been independently verified. Glimpse is fine-tuned on an open-weight model. Elorian states its small team uses just hundreds of GPUs. The model’s API charges $0.60 per million input tokens and $3 per million output tokens, both 94% lower than Astra’s rates. Currently, the model is available as a research preview API, with its weights not yet made public. Elorian previously closed a $55 million seed round, valuing the company at $300 million; investors include Nvidia and former Google Chief Scientist Jeff Dean.