The creator of Ghostty has added an activity view to its new terminal Rex, which centrally displays agent status.

Bitcoin’s weekly close held the key support level of $82,500, temporarily avoiding a fall back into its long-term trading range of $60,000 to $80,000. BTC settled at around $83,700 on Sunday, down roughly 3.2% for the week. Traders are closely monitoring this week’s U.S. inflation data to assess future market trends. September’s CPI and PPI figures will be released on Wednesday and Thursday respectively. The market expects the year-over-year CPI rise to climb from 3.4% to 3.7%, while the month-over-month PPI increase is projected to go from 0.2% to 0.3%. Lower-than-expected inflation could ease concerns over further Federal Reserve interest rate hikes, providing support for risk assets including Bitcoin, stocks and gold; higher-than-expected inflation, meanwhile, may intensify market pressure. On the rate front, CME FedWatch data shows the market assigns an 18.8% probability of a 25-basis-point rate hike by the Fed on October 28, and a 70.6% chance of a 25-basis-point hike in December. Separately, U.S. 10-year and 30-year Treasury yields stand at 5.24% and 5.60% respectively. Additionally, WTI crude oil rebounded to around $91.50 per barrel on Monday, up roughly 2% from its intraday low. Traders note that if Bitcoin can sustain its hold on the $82,500 support level, it may continue its rally; a break below this level, however, could expose it to renewed downward pressure.

Deutsche Bank raises Microsoft (MSFT.O) price target from $550 to $600.

The Financial Stability Board (FSB) warned in its latest report that global stablecoins have not yet withstood severe market stress or bankruptcy scenarios, as national regulatory frameworks advance at uneven paces and cross-border collaboration remains in its early stages. The FSB called for strengthened cross-border information sharing, risk monitoring, and advance contingency plans for when stablecoin issuers face distress or collapse. The FSB specifically noted that multi-issuer structures could amplify cross-border risks: affiliated companies issue the same stablecoin across different jurisdictions and hold separate reserve assets, but users may redeem from any issuer in the group. If redemption demand concentrates on one issuer while reserve assets are held in other regions, fund transfers during a crisis could be blocked, and creditor claims plus applicable bankruptcy laws could be hard to determine if an issuer goes bankrupt. Additionally, stablecoins’ 24/7 trading does not align with the trading hours of their underlying reserve assets, which may lead to redemption pressure failing to be met via timely asset sales and trigger depegging. Forced sales of reserve assets and concentrated bank deposit withdrawals could also transmit shocks to traditional financial markets. The FSB also pointed out that the anonymity of crypto transactions may make it harder to identify holders’ locations and their applicable regulatory jurisdictions. Data shows that the total market capitalization of stablecoins has risen from roughly $206 billion at the start of 2025 to around $306 billion, with the majority being dollar-pegged stablecoins. The FSB’s recommendations are not legally binding, but its members typically commit to implementing the relevant standards domestically. The report is released ahead of the G20 finance ministers and central bank governors’ meeting scheduled for October 15.

Beating AI Express: AI research institute PhAI Labs has launched RSI-Anything, a framework enabling AI agents to continuously accumulate experience from real tasks and improve their performance. The framework supports use cases including code debugging, computer operations, and scientific research. Its core design equips agents with a "world model" that predicts the outcomes of different actions. The agent first proposes multiple solutions; the world model evaluates which is most likely to succeed, after which the agent selects that solution for execution. The system then compares predictions with actual results to refine the agent’s operational strategy. If predictions are inaccurate, the world model is also updated. Only after the new version passes tests is it retained for subsequent tasks. Take bug fixing as an example: the code-generating model produces four modification proposals at once. RSI-Anything first predicts which is most likely to pass tests, selects only one to run, then uses the test result to improve future judgments. This approach allocates limited test attempts to more promising solutions. In 200 CodeContests problems, the code-writing model ReasonFlux-Coder-7B remained unchanged throughout, with roughly the same number of test attempts. Using a traditional static verifier, the bug-fixing pass rate stood at 26.3%. After introducing the world model for screening, the rate rose to 29.2%; further updating the world model with real feedback pushed it to 31.2%. The team also applied this method to computer operations and scientific experiments. In a lab planning test, a 4B model was updated five consecutive times, with its planning score rising from 21.80% to 46.01%. The project code is now open-source. Currently, separate agents and world models are used for different tasks. The paper acknowledges that whether the execution strategy and world model can improve together long-term on the same task requires additional experimental verification.

Binance founder CZ posted on X that he will never charge for brief meetings, and will refuse to meet anyone who pays for a chance to see him. He also reminded users not to pay any middlemen or scammers. CZ added that while it’s impossible to meet everyone, he often arranges several 5-minute one-on-one sessions for community supporters during BBW events, and expressed hope to meet community members in Thailand in November. Earlier, X user @thevoid_bnb discussed the value of brief exchanges with industry figures in a post, citing Justin Sun’s $4.6 million lunch with Warren Buffett as an example, and asking what questions people would raise if they had a 5 to 10 minute one-on-one chat with CZ. CZ later responded to the post, clearly stating he would not offer paid brief meetings.

UBS said that despite a stronger U.S. dollar and rising U.S. real rates weighing on gold prices, and short-term speculative funds cutting net long positions in gold futures and options, gold ETFs have continued to attract capital inflows recently, signaling that medium and long-term allocation demand remains intact. The bank projects gold prices could find support near $4,000 per ounce, and rise to $4,600 in December 2026, $5,000 in March 2027, $5,200 in June 2027, and hit $5,400 in September 2027. UBS attributes gold’s support to sustained global central bank gold purchases, robust demand from China, and potential physical gold demand spurred by India’s festival season. Additionally, investor concerns over public debt, fiscal risks, and the long-term outlook for the U.S. dollar may also enhance gold’s allocation appeal. That said, if the U.S. dollar and real yields continue to strengthen, gold prices could still see further short-term pullbacks. For its part, Credit Agricole forecasts gold prices will reach $5,000 per ounce in December 2026. Both institutions agree that central bank reserve diversification and market worries over the Federal Reserve’s policy being impacted by fiscal pressures could drive gold’s medium and long-term gains.