GateUser-94818fd0

Netomi founder Puneet Mehta states that by 2030, the customer experience industry will reach $5 trillion, driven by demand for stablecoins and blockchain payment infrastructure rather than capital from the crypto sector; the company's annual expenditure on customer experience-related knowledge work is about $500 billion. As AI expands from customer service to sales and other areas, market opportunities are expected to grow tenfold. He believes AI and crypto are complementary, with AI agents requiring faster financial infrastructure, which could become a driving force for stablecoin adoption.

ETH was dead at $200.
ETH was dead at $800.
ETH was dead at $1,400.
Now, $ETH is dead at $1,600.
You know what's coming next.

Crypto news, according to Lookonchain monitoring, Huang Licheng experienced 10 liquidation events within 8 hours, with his account balance dropping to $52,000. He then increased his long position to 1,075 ETH, which is worth approximately $1.71 million at the current price, with a new liquidation price of $1,560.81.

ME News message. April 19 (UTC+8). Forecast market data shows that within 24 hours, multiple addresses were concentrated in participating in the event “Will the U.S. strike Iran before February 22, 2026?” and selected the outcome “Yes.” The total bet amount is approximately $2098.5k, indicating that this result received significant funding attention in a short period.

CryptoWorld News reports that Swapzone has announced the addition of real-time KYC and execution data in its pre-trade interface, allowing users to assess a provider’s operational risk before submitting funds. This update addresses structural issues identified through independent research quantification. According to Bitcoin.com’s speed benchmark: the Non-Custodial Exchange Comparison 2026 report, based on Swapzone data covering 150,000 transactions, execution speed has become a key factor in value preservation. The study found that performance gaps between leading platforms and median platforms can reach up to 45 times. For example, a $1,000 USDT to ETH swap with an industry median execution time of about 45 minutes results in a 3.2% loss, while platforms with execution times under 1 minute only lose 0.1%. Each additional minute of execution time is expected to reduce the value of every $1,000 swap by $2–5. Swapzone’s aggregation architecture, by drawing from 18

Chainalysis reports that the stolen wallet from THORChain transferred funds over several weeks through Monero, Hyperliquid, and THORChain. In April, funds were deposited into Hyperliquid via a privacy bridge, then transferred to USDC on Arbitrum before bridging to Ethereum, with some ETH entering THORChain to become new RUNE nodes. The attacker split RUNE into four routes back to Ethereum, one of which directly sent 8 ETH to the final wallet about 40 minutes before the attack. The other three routes flowed funds in reverse through Hyperliquid/Monero back to Monero. As of Friday, the funds have not been moved, but the cross-chain money laundering capability has been demonstrated, with the Hyperliquid to Monero path potentially being the next step.

ByteDance and Qualcomm have reached a procurement and contract manufacturing agreement for millions of custom AI ASIC chips, which will be mass-produced by manufacturing partners such as TSMC and used in AI systems like Doubao. To address geopolitical scrutiny, the design adopts pixel-level alignment compliance avoidance strategies to ensure it stays within legal thresholds. This move alleviates dependence on NVIDIA GPUs and marks Qualcomm's entry into the AI infrastructure market, with ByteDance becoming one of the first major clients. ByteDance's AI infrastructure budget has been increased to approximately 200 billion RMB.

AIMPACT proposes a systematic approach to convert post-training inference models into Olympic-level problem solvers, in three steps: supervised fine-tuning with an inverse perplexity curriculum to instill proof search and self-checking; then expanded through two-stage reinforcement learning; and finally, perform scaling improvements during testing. Applied to the 30B-A3B backbone, it uses approximately 340k sub-8K trajectories for supervised fine-tuning, followed by 200 steps of RL, to obtain SU-01. This model can carry out stable reasoning on challenging problems; its trajectories exceed 100k tokens. In competitions such as IMO, USAMO, and IPhO, it achieves gold-medal-level performance, and also demonstrates scientific reasoning generalization across fields beyond mathematics and physics.

Princeton computer science professor Arvind Narayanan stated at the Stanford Digital Economy Lab seminar that AI will automate a large amount of cognitive labor, but downstream capacity bottlenecks will gradually emerge over the next few decades. He criticized the evidence infrastructure for over-focusing on capability dimensions and introduced his team's work in measuring diffusion characteristics, such as open-world evaluation and viewing AI reliability as orthogonal to capability. He proposed a future-oriented cognitive labor automation agenda, predicting workforce demand, institutional risks, and new social ethical challenges, advocating for the development of contextual awareness and a dual-track approach to predicting new equilibria.

Coinbase's Ethereum Layer 2 network Base launches AI tool Base MCP, which can connect users' Base wallets to AI clients such as ChatGPT, Claude, and Cursor through the Model Context Protocol (MCP). Users can perform operations like transferring funds, token swaps, balance inquiries, transaction history viewing, and DeFi interactions using natural language commands. The initial integrated protocols include Uniswap, Morpho, Moonwell, and Avantis. Base states that it is betting on AI agents and conversational interfaces becoming important gateways for discovering and interacting with on-chain applications in the future. (CoinDesk)