AI learned to hack servers and copy itself across the network — study findings
A study by Palisade Research, published on May 7, 2026, recorded the first documented case of language models autonomously hacking systems and fully replicating themselves on compromised machines. What exactly did the agents achieve? According to the organization's report, agents based on models like Qwen3.5-122B-A10B autonomously discovered vulnerabilities in web applications — hash bypass, server-side template injection, SQL injections, and broken access control — extracted credentials, and deployed a full copy of themselves on the hacked hosts with everything needed to operate. This fundamentally distinguishes the experiment from previous works, where agents were given prior access to the target.
Island with a population of 15k nets $85 million a year from the AI boom
February 2026 was a record month for Anguilla: this tiny Caribbean territory raked in $32.8 million Eastern Caribbean dollars in just one month—thanks to the .ai domain. The story of this island with a population of 15-16 thousand continues to go viral on social media as a prime example of how a two-letter acronym has transformed the economy of an entire nation.
AI Agents in Finance: What They Can Do, What They Don't Understand, and Where They Break Down
Autonomous AI systems are gradually entering the financial processes of businesses. But before trusting an agent with the budget or trading strategy, it's worth understanding how they work — and why autonomy doesn't equal reliability. The agent doesn't know what it's doing. And that's not a metaphor. One of the most common myths about AI is that systems 'think' or 'make decisions.' In practice, language models operate as statistical predictors: they analyze patterns in data and generate the most probabilistically suitable next move.
Stablecoins: what's behind them and why 'stable' isn't always an accurate term
The stablecoin market has surpassed $200 billion. Major banks, payment networks, and governments are discussing their integration. However, fundamental questions about how they work and the risks for users still lack clear answers for most market participants. Why stablecoins aren't synonymous with reliability
Solana Strengthens Defense Against Quantum Threats, Ethereum L2 Faces Criticism
The topic of quantum computing is back in the spotlight of the crypto market. This was sparked by statements from Solana co-founder Anatoly Yakovenko, who questioned the resilience of Ethereum Layer 2 against future quantum threats while also discussing Solana's advancements towards post-quantum cryptography.
Privacy and transparency in blockchain do not contradict each other. How it works
At Consensus Miami, representatives from Moody’s and ChangeNOW explained why the conflict between user anonymity and regulatory demands is not a dead-end but rather an engineering challenge with workable solutions. False dilemma Public blockchains are transparent by nature: every transaction is recorded, accessible, and traceable. This makes them suitable for auditing and monitoring — but at the same time, it exposes users' financial activity to an extent that clashes with the fundamental concept of privacy.
Kraken buys Reap for $600 million and doubles down on stablecoins
Kraken, through its parent company Payward, is diving deeper into the payment infrastructure market. The company has agreed to acquire Hong Kong's Reap Technologies for up to $600 million. This deal will be one of the largest for the group and will strengthen its position in the stablecoin payment segment and B2B services. The market is gradually shifting from regular crypto trading to financial infrastructure. Exchanges are increasingly building systems for international transfers, settlements, and corporate payments.
JP Morgan thinks that oil stocks are currently overvalued by about $40. The bank attributes this mainly to the tension between the USA and Iran, which continues to pump up oil prices. This week, earnings reports for Q1 showed three major oil companies. The market reacted differently to them. Some investors are betting on a more stable business, others still believe in production growth, while certain stocks could seriously tank if oil starts losing its geopolitical support.
Funds are diversifying assets while Strategy considers its first BTC short since 2020.
The study included 26 fund managers with assets of around $1.3 trillion. For most crypto, it's now more about risk diversification and responding to client demand than trying to make a quick profit. They accounted for 63% of the responses from survey participants. For comparison: just two years ago, this metric was only 36%.
Strategy may start selling Bitcoin for the first time in 5 years
Michael Saylor has hinted at selling Bitcoin — and this marks the first instance where the Executive Chairman of Strategy has publicly deviated from his long-standing position of 'never selling'. Saylor mentioned this during a Q1 earnings call with investors. According to him, the company might liquidate some Bitcoin to pay dividends — not out of financial necessity, but to 'immunize the market' against panic and affirm the resilience of the business. 'We’ll probably sell a bit of Bitcoin to fund dividends, just to instill confidence in the market, just to send the signal: we did it,' Saylor stated. Market participants, in his view, will be reassured that 'the company is fine, Bitcoin is fine, the industry is fine, the world hasn't collapsed.'
Make TON Great Again: Pavel Durov pumped Toncoin by 65%
TON has shot up sharply over the last three days: the coin gained 65% — climbing from around $1.36 to $2.24. This movement is one of the most significant rises in the past year: on the daily candlestick chart, the bullish candle stands out against a multi-month downtrend that has been in place since summer 2025.
a16z has raised $2.2 billion for its fifth crypto fund. This is one of the largest specialized crypto funds launched this year. Institutional money keeps flowing into blockchain startups even in a calm market, despite the relatively stable conditions.
Why did Intel (INTC) shares suddenly spike over 17%, and how much has Trump already made off this?
Intel (INTC) shares are back in the spotlight on Wall Street after a massive rally, with the stock surging over 17% in a single day. On Tuesday, the shares jumped 13% to hit a record $110, and after-hours trading saw an additional increase of 4.76%. The catalyst for this movement emerged quickly. Traders are betting that Intel could play a bigger role in collaboration with Apple (AAPL), especially as chip production in the U.S. becomes increasingly vital for major tech companies.
Unity has launched an AI agent for game development without a single line of code
Unity has launched an open beta for its suite of AI tools, Unity AI – now developers can create game scenes, assets, and entire projects through simple text prompts without writing a single line of code. The toolkit is available for Unity 6 and above starting from early May 2026. The centerpiece of the toolkit is the AI agent Assistant, integrated right into the editor. It’s trained on over 20 years of best practices in Unity and understands the full context of a project: scene hierarchy, settings, packages, code, asset library, target platforms, and inspector data. The Assistant operates in three modes:
U.S. Banks Oppose CLARITY Act: Stablecoins Should Not Generate Yield
The largest banking associations in the U.S. expressed dissatisfaction with the wording of the CLARITY Act regarding the regulation of yield on stablecoins, stating that the document does not provide adequate protection for bank deposits. In a joint statement released on Monday, the American Bankers Association, Banking Policy Institute, Consumer Bankers Association, Financial Services Forum, and Independent Community Bankers Association acknowledged that Senators Thom Tillis and Angela Alsobrooks are "aiming for the right policy goal" in banning yield on stablecoins. "It is crucial for Congress to make the right decision," the statement said.
DEXs are losing volume in April, outflows are hitting the entire market
In April, DEX activity took another hit, and this is starting to look like more than just a short-term dip. The downtrend has been dragging since October 2025. It's not just the volumes that are tanking — liquidity is also drying up, and it’s flowing out of both spot and derivatives. According to DeFi Llama, the total trading volume for the month hit $166.78 billion. That's the lowest since last August. Compared to the peak in October, the market has already retraced about 59%.
Bitcoin dominance is decreasing, Ethereum could trigger an alt season
Ethereum is currently compressing under the $2,300 level, while Bitcoin dominance is nearing the formation of a monthly 'death cross'—a signal that has only occurred twice in history. In the past, such a combination often marked the starting point for an alt season. These two charts are moving almost in sync. ETH remains the top altcoin, and when it starts to gain momentum against Bitcoin, capital gradually flows out of BTC. As a result, dominance decreases, and this kind of setup has previously often preceded a shift in favor of altcoins.
Iran reported that Trump rejected the plan to open the Strait of Hormuz
Iran stated that Trump rejected the proposal in which Tehran was ready to open the Strait of Hormuz, lift the maritime blockade, and postpone nuclear talks for later. According to sources, the strategy was to play it safe: first, halt the conflict and sort out the shipping security, and only then circle back to the nuclear issue.
Physical AI: when intelligence stops being just digital
Physical AI is becoming the next big phase in the development of artificial intelligence (AI) — not because humanoid robots have suddenly become a reality, but because intelligence is increasingly making its way into the physical world. Until recently, the talk about AI revolved around digital systems: models generating text, summarizing information, writing code, and answering questions. This wave has truly changed a lot. But the next stage will be defined not by what systems can say, but by what they can actually do. That's the industry's belief — and it's hard to argue with that.