Ice Lounge Media

Ice Lounge Media

Slovenia’s finance ministry floats 25% tax on crypto transactions

Slovenia’s Finance Ministry is considering a possible 25% tax on crypto trading profits for residents in the country under a new draft law now open for public consultation. 

The bill proposes to tax traders when they sell their cryptocurrency for fiat or pay for goods and services, but crypto-to-crypto and transfers between wallets owned by the same user will be exempt, Slovenia’s Finance Ministry said in an April 17 statement.

Under the proposed legislation, crypto tax will be aligned with existing tax laws. Slovenia taxpayers will be required to keep a record of all their transactions for annual tax returns. The tax base would be calculated on profits by subtracting the purchase price from the sale price. 

In a statement to the Slovenia Times, finance minister Klemen Boštjančič said it’s unreasonable that crypto trading for individuals isn’t currently taxed in the country. 

“The goal of taxation of crypto assets is not to generate tax revenue, but we find it illogical and unreasonable that one of the most speculative financial instruments is not taxed at all,” he said in a statement translated from Slovenian.

New tax could stifle crypto in Slovenia, lawmaker says 

Jernej Vrtovec, a member of Slovenia’s national assembly and New Slovenia opposition party, slammed the proposal in an April 16 statement to X, arguing it could stifle crypto growth in the country. 

“Slovenia has the opportunity to become a crypto-friendly country, but with the government’s proposals, we will miss the train again,” he said in a post also translated from Slovenian.

“With excessive taxation, we will once again see young people and capital fleeing abroad. Taxes should encourage, not stifle.” 

Slovenia’s finance ministry floats 25% tax on crypto transactions
Source: Jernej Vrtovec

The proposal is open to public consultation until May 5. If Slovenian lawmakers pass the bill, it will go into effect on Jan. 1, 2026. 

Slovenia introduced a 10% tax on crypto withdrawals and payments in 2023, but capital gains from occasional crypto trading are not taxed, according to the crypto tax platform Token Tax. 

Related: NFT trader faces prison for $13M tax fraud on CryptoPunk profits

Crypto activity can also currently be exempt from tax if it’s considered a hobby. Business activity, such as mining or staking, is subject to income tax. 

A previous bill proposed in April 2022 planned to levy a 5% tax on profits over 10,000 euros ($11,372), but it was never passed into law. 

Slovenia issued the first digital sovereign bond in the European Union on July 25 last year. It had a nominal size of 30 million euros ($32.5 million) with a 3.65% coupon and a maturity date of Nov. 25 that year. 

The number of crypto users in Slovenia is projected to reach roughly 98,000 in 2025, according to online data platform Statista, with a penetration rate of 4.6% among its population of 2.12 million people. While the projected revenue for the country’s crypto market is slated to hit $2.8 million. 

Magazine: How crypto laws are changing across the world in 2025

Read more

Bitcoin dip buyers nibble at BTC range lows but are risk off until $90K becomes support

Bitcoin’s (BTC) realized market cap reached a new all-time high of $872 billion, but data from Glassnode reflects investors’ lack of enthusiasm at BTC’s current price levels.

In a recent X post, the analytics platform pointed out that despite the realized cap milestone, the monthly growth rate of the metric has dropped to 0.9% month over month, which implied a risk-off sentiment in the market.

Coinbase, Cryptocurrencies, Bitcoin Price, Markets, Price Analysis, Market Analysis
Bitcoin realized cap net position. Source: X.com

Realized cap measures the total value of all Bitcoin at the price they last moved, reflecting the actual capital invested, providing insight into Bitcoin’s economic activity. A slowing growth rate highlights a positive but reduced capital inflow, suggesting fewer new investors or less activity from current holders.

Additionally, Glassnode’s realized profit and loss chart recently exhibited a sharp decline of 40%, which signals high profit-taking or loss realization. The data platform explained,

“This suggests saturation in investor activity and often precedes a consolidation phase as the market searches for a new equilibrium.”

While new investors remained sidelined, existing investors are probably adopting a cautious approach due to the short-term holder’s realized price. Data from CryptoQuant suggested that the current short-term realized price is $91,600. With BTC currently consolidating under the threshold, it implies short-term holders are underwater, which can increase selling pressure if they sell to cut their losses.

Coinbase, Cryptocurrencies, Bitcoin Price, Markets, Price Analysis, Market Analysis
Bitcoin short-term holders’ price and MVRV. Source: CryptoQuant

Similarly, Bitcoin’s short-term holder market value to realized value remained below 1, a level historically associated with buying opportunities and further proof that short-term holders are at a loss.

Related: Bitcoin US vs. offshore exchange ratio flashes bullish signal, hinting at BTC price highs in 2025

Bitcoin chops between US and Korean traders

Data shows a sentiment divergence between Bitcoin traders in the US and Korea. The Coinbase premium, reflecting US trading, recently spiked, signaling strong US demand and potential Bitcoin price gains.

Conversely, the Kimchi premium index fell during the correction, indicating lagging retail engagement among Korea-based traders.

This particular uneven demand is reflected in Bitcoin’s recent price action. The chart shows that Bitcoin’s price has oscillated between a tight range of $85,440-$82,750 since April 11. On the 4-hour chart, BTC has retained support from the 50-day, 100-day, and 200-day moving averages, but on the 1-day chart, these indicators are putting resistance on the bullish structure.

Coinbase, Cryptocurrencies, Bitcoin Price, Markets, Price Analysis, Market Analysis
Bitcoin 4-hour chart. Source: Cointelegraph/TradingView

Related: Bitcoin online chatter flips bullish as price chops at $85K: Santiment

This article does not contain investment advice or recommendations. Every investment and trading move involves risk, and readers should conduct their own research when making a decision.

Read more

Kyrgyzstan’s president signs CBDC law giving ‘digital som’ legal status

Kyrgyzstan President Sadyr Zhaparov has signed a constitutional law authorizing the launch of a central bank digital currency pilot project while also giving the “digital som” — the national currency in digital form — legal tender status.

The law gives the National Bank of the Kyrgyz Republic the exclusive right to issue the digital som, establish the rules for its issuance and circulation, and oversee the platform on which the national currency will operate, Kyrgyzstan’s presidential office said on April 17.

However, a final decision on whether to officially issue the CBDC is not expected until the end of 2026, local outlet Trend News Agency reported in December.

If the central bank decides to adopt the digital som, it would also need to outline cryptographic protection measures to ensure the digital som remains secure and isn’t used for fraudulent transactions.

Testing of the digital som platform is expected to take place sometime this year.

Zhaparov’s sign-off comes nearly a month after Kyrgyzstan’s parliament, the Jogorku Kenesh, approved the amendment to Kyrgyzstan’s constitutional law on March 18.

CBDCs continue to be heavily criticized by some members of the crypto community, flagging concerns that they could undermine financial privacy and enable excessive government oversight, among other things.

While 115 nations have initiated CBDC projects, only four CBDCs have officially launched — the Bahamas Sand Dollar, Nigeria’s e-Naira, Zimbabwe’s ZiG and Jamaica’s JAM-DEX, data from cbdctracker.org shows.

Over 90 CBDC projects are yet to move past the research stage.

Kyrgyzstan continues to make moves in crypto

Earlier this month, former Binance CEO Changpeng “CZ” Zhao said he would begin advising Kyrgyzstan on blockchain and crypto-related regulation after signing a memorandum of understanding with the country’s foreign investment agency.

Zhaparov said the initiative would assist with the growth of the economy and the security of virtual assets, “generating new opportunities for businesses and society as a whole.”

Kyrgyzstan’s president signs CBDC law giving ‘digital som’ legal status
Source: Sadyr Zhaparov

Related: Bitcoin price levels to watch as Fed rate cut hopes fade

The mountainous, land-locked country is considered well-suited for crypto mining operations due to its abundant renewable energy resources, much of which is underutilized.

Over 30% of Kyrgyzstan’s total energy supply comes from hydroelectric power plants, but only 10% of the country’s potential hydropower has been tapped, according to a report by the International Energy Agency.

Magazine: Your AI ‘digital twin’ can take meetings and comfort your loved ones

Read more

OpenAI sought Anysphere deal before turning its sights on WindSurf

OpenAI was reportedly in talks to buy Anysphere, the company that produces the Cursor AI coding assistant, before entering into talks with rival company WindSurf.

According to CNBC, OpenAI approached Anysphere in 2024 and again in 2025, but talks stalled both times. Failing to arrive at a deal led OpenAI to look elsewhere for potential acquisitions.

Sources familiar with the deal also say OpenAI is prepared to pay $3 billion to purchase WindSurf, which would make it the company’s largest corporate acquisition to date.

OpenAI
An example of OpenAI’s ChatGPT producing computer code through simple text prompts. Source: ChatGPT

OpenAI’s attempted acquisition of an AI coding assistant company follows the release of DeepSeek R1 in January 2025, which shattered long-held assumptions about artificial intelligence.

DeepSeek was reportedly trained at a fraction of the cost of leading AI models while delivering comparable performance — challenging the belief that scaling requires massive computing power, rattling financial markets, and raising questions about the billions spent by US AI giants.

Related: OpenAI to release its first ‘open’ language model since GPT-2 in 2019

OpenAI inches toward profitability but cheaper competitors still a challenge

OpenAI expects to triple its revenue in 2025 to approximately $12.7 billion by selling paid subscriptions for its leading AI models to individuals and businesses.

The company surpassed 1 million premium business subscribers in September 2024. However, OpenAI CEO Sam Altman said the AI giant might not be profitable until 2029.

According to Altman, OpenAI needs revenues of approximately $125 billion to turn a profit on its capital-intensive business.

In February 2025, Altman said that AI development costs were dropping dramatically. “The cost to use a given level of AI falls about 10x every 12 months,” the CEO wrote in a Feb. 9 blog post.

Despite this, high costs and centralization issues continue to plague large-scale corporate AI developers, who must compete with more nimble open-source counterparts.

Dr. Ala Shaabana — co-founder of the OpenTensor Foundation — recently told Cointelegraph that the release of DeepSeek solidified open-source AI as a serious contender against centralized AI systems.

Shaabana added that the lower cost of open-source systems proves that AI does not need billions of dollars to scale or achieve high-performance benchmarks.

Magazine: 9 curious things about DeepSeek R1: AI Eye

Read more
Expense management startup Ramp is being considered for a charge card pilot program by the U.S. government’s General Services Administration, the company confirmed to TechCrunch on Thursday. The government’s internal expense card program, dubbed SmartPay, is a $700 billion program. It is estimated that the charge card pilot program contract for which Ramp is being […]
Read more
In a bid to more aggressively compete with rival AI companies like Google, OpenAI is launching Flex processing, an API option that provides lower AI model usage prices in exchange for slower response times and “occasional resource unavailability.” Flex processing, which is available in beta for OpenAI’s recently released o3 and o4-mini reasoning models, is […]
Read more

Google DeepMind’s latest update to a top Gemini AI model includes a dial to control how much the system “thinks” through a response. The new feature is ostensibly designed to save money for developers, but it also concedes a problem: Reasoning models, the tech world’s new obsession, are prone to overthinking, burning money and energy in the process.

Since 2019, there have been a couple of tried and true ways to make an AI model more powerful. One was to make it bigger by using more training data, and the other was to give it better feedback on what constitutes a good answer. But toward the end of last year, Google DeepMind and other AI companies turned to a third method: reasoning.

“We’ve been really pushing on ‘thinking,’” says Jack Rae, a principal research scientist at DeepMind. Such models, which are built to work through problems logically and spend more time arriving at an answer, rose to prominence earlier this year with the launch of the DeepSeek R1 model. They’re attractive to AI companies because they can make an existing model better by training it to approach a problem pragmatically. That way, the companies can avoid having to build a new model from scratch. 

When the AI model dedicates more time (and energy) to a query, it costs more to run. Leaderboards of reasoning models show that one task can cost upwards of $200 to complete. The promise is that this extra time and money help reasoning models do better at handling challenging tasks, like analyzing code or gathering information from lots of documents. 

“The more you can iterate over certain hypotheses and thoughts,” says Google DeepMind chief technical officer Koray Kavukcuoglu, the more “it’s going to find the right thing.”

This isn’t true in all cases, though. “The model overthinks,” says Tulsee Doshi, who leads the product team at Gemini, referring specifically to Gemini Flash 2.5, the model released today that includes a slider for developers to dial back how much it thinks. “For simple prompts, the model does think more than it needs to.” 

When a model spends longer than necessary on a problem, it makes the model expensive to run for developers and worsens AI’s environmental footprint.

Nathan Habib, an engineer at Hugging Face who has studied the proliferation of such reasoning models, says overthinking is abundant. In the rush to show off smarter AI, companies are reaching for reasoning models like hammers even where there’s no nail in sight, Habib says. Indeed, when OpenAI announced a new model in February, it said it would be the company’s last nonreasoning model. 

The performance gain is “undeniable” for certain tasks, Habib says, but not for many others where people normally use AI. Even when reasoning is used for the right problem, things can go awry. Habib showed me an example of a leading reasoning model that was asked to work through an organic chemistry problem. It started out okay, but halfway through its reasoning process the model’s responses started resembling a meltdown: It sputtered “Wait, but …” hundreds of times. It ended up taking far longer than a nonreasoning model would spend on one task. Kate Olszewska, who works on evaluating Gemini models at DeepMind, says Google’s models can also get stuck in loops.

Google’s new “reasoning” dial is one attempt to solve that problem. For now, it’s built not for the consumer version of Gemini but for developers who are making apps. Developers can set a budget for how much computing power the model should spend on a certain problem, the idea being to turn down the dial if the task shouldn’t involve much reasoning at all. Outputs from the model are about six times more expensive to generate when reasoning is turned on.

Another reason for this flexibility is that it’s not yet clear when more reasoning will be required to get a better answer.

“It’s really hard to draw a boundary on, like, what’s the perfect task right now for thinking?” Rae says. 

Obvious tasks include coding (developers might paste hundreds of lines of code into the model and then ask for help), or generating expert-level research reports. The dial would be turned way up for these, and developers might find the expense worth it. But more testing and feedback from developers will be needed to find out when medium or low settings are good enough.

Habib says the amount of investment in reasoning models is a sign that the old paradigm for how to make models better is changing. “Scaling laws are being replaced,” he says. 

Instead, companies are betting that the best responses will come from longer thinking times rather than bigger models. It’s been clear for several years that AI companies are spending more money on inferencing—when models are actually “pinged” to generate an answer for something—than on training, and this spending will accelerate as reasoning models take off. Inferencing is also responsible for a growing share of emissions.

(While on the subject of models that “reason” or “think”: an AI model cannot perform these acts in the way we normally use such words when talking about humans. I asked Rae why the company uses anthropomorphic language like this. “It’s allowed us to have a simple name,” he says, “and people have an intuitive sense of what it should mean.” Kavukcuoglu says that Google is not trying to mimic any particular human cognitive process in its models.)

Even if reasoning models continue to dominate, Google DeepMind isn’t the only game in town. When the results from DeepSeek began circulating in December and January, it triggered a nearly $1 trillion dip in the stock market because it promised that powerful reasoning models could be had for cheap. The model is referred to as “open weight”—in other words, its internal settings, called weights, are made publicly available, allowing developers to run it on their own rather than paying to access proprietary models from Google or OpenAI. (The term “open source” is reserved for models that disclose the data they were trained on.) 

So why use proprietary models from Google when open ones like DeepSeek are performing so well? Kavukcuoglu says that coding, math, and finance are cases where “there’s high expectation from the model to be very accurate, to be very precise, and to be able to understand really complex situations,” and he expects models that deliver on that, open or not, to win out. In DeepMind’s view, this reasoning will be the foundation of future AI models that act on your behalf and solve problems for you.

“Reasoning is the key capability that builds up intelligence,” he says. “The moment the model starts thinking, the agency of the model has started.”

This story was updated to clarify the problem of “overthinking.

Read more
1 10 11 12 13 14 2,656