Latest news

The Social Insurance Institution of Finland, known as Kela, has suffered a massive technical failure starting in the early hours of Thursday, April 23, 2026.

Kela is a government agency responsible for settling benefits under national social security programs, including pensions, unemployment benefits, and student grants, making its digital infrastructure critical for millions of residents. The outage has effectively paralyzed almost all of the agency's customer services, including its online portal, telephone services, and remote appointment systems.

Kela's official website is currently offline, replaced by a temporary emergency site to provide basic information. The primary digital service portal, OmaKela, is also completely inaccessible, preventing users from submitting new benefit applications or checking the status of their payments. According to the agency, remote customer service remains impossible while the systems are down.

Despite the severity of the situation, Kela's IT Director Jukka Melanen has stated that there are currently no indications that the outage was caused by a distributed denial-of-service (DDoS) attack or any other form of cyber interference. The agency's IT department is working closely with its technical partners to identify the root cause and restore services as quickly as possible.

The European Commission has issued its preliminary findings, according to which search engine giant Google must open its search data for use by third parties.

The proposed measures are based on the EU's Digital Markets Act (DMA), and their goal is to dismantle Google's dominant position in the market. The Commission demands that Google must share data, such as search result rankings, search queries, and click and view counts, on fair and equitable terms.

According to the proposal, the recipients of the data would not only be traditional competing search engines but also AI-powered chatbots that utilize search functionalities.

The Commission believes that access to this information is critical for other players to optimize their own services and offer a real challenge to Google.

Commission Vice-President Henna Virkkunen emphasized that search engines must be able to innovate and keep up with users' changing needs, especially now that AI services are becoming increasingly integrated into searches.

Google has announced that it will fight against the proposals.

The hated so-called Chat Control bill was supposed to fade into the mists of history when the privacy-infringing parts of the legislative initiative failed in the spring of 2026 in a European Union vote.

Chat Control is thus a bill that has been attempted to be pushed through the European Union since 2023 (PDF).

The core of the bill is that all message encryption should be banned throughout the entire European Union.

This means WhatsApp messages could no longer be encrypted in such a way that they could be deemed secure. The same would apply to emails, Signal, and any other message traffic.

The initiative is justified by its supporters, like practically all other ideas that strongly infringe on people's fundamental rights, with the so-called "think of the children...!" argument.

Samsung and furniture giant IKEA have announced a major collaboration that makes building a smart home more affordable and straightforward than before.

Samsung's SmartThings platform will from now on directly support 25 new IKEA devices that use the Matter-over-Thread standard. IKEA released a bunch of Matter-compatible smart devices at the end of last year.

In practice, this means that consumers can connect IKEA's popular smart bulbs, sockets and sensors directly to Samsung's smart home hub without needing IKEA's own Dirigera bridge.

Previously, cross-using these two systems often required both manufacturers' own hubs, which increased both costs and technical complexity.

Technically, the change relies heavily on the Matter-over-Thread combination, which is one of the most significant developments in the smart home world in years.

Matter is, so to speak, a common language between manufacturers that ensures compatibility between devices from different brands, while Thread is a low-power mesh networking technology designed specifically for the needs of small smart home devices.

Amazon is deepening its collaboration with the rapidly growing AI company Anthropic through a 25-billion-dollar investment and a cloud services commitment worth over 100 billion dollars.

Amazon will immediately invest 5 billion dollars in Anthropic and has committed to investing an additional 20 billion dollars in the company if predefined commercial and technical targets are met. The new funding comes on top of two previous 4-billion-dollar rounds in 2023 and 2024, bringing Amazon's total potential investment from 8 billion to 33 billion dollars - if all targets are achieved.

In return, Anthropic commits to using Amazon's cloud services and, in particular, the company's custom Trainium chips to train and run its AI models over the next ten years. As part of the capacity agreement, Anthropic secures up to 5 gigawatts of current and future chip capacity - a scale that the companies compare to the electricity production of five large nuclear power plants.

Over the past six months, Anthropic has faced what is often called a positive problem, as the popularity of its AI solutions, especially among enterprise users, has exploded thanks to the growing success of Claude Code. Anthropic has been forced to tighten usage limits even for paying customers to ensure its server capacity can keep up with demand.

Apple has elected the company's next CEO. Tim Cook, Apple's current CEO and successor to Steve Jobs, will become the company's executive chairman.

Apple's upcoming new CEO is John Ternus who has worked for Apple since 2001 and is currently the senior vice president of Hardware Engineering.

In company's official press release, Tim Cook described Mr. Ternus like this:

John Ternus has the mind of an engineer, the soul of an innovator, and the heart to lead with integrity and with honor. He is a visionary whose contributions to Apple over 25 years are already too numerous to count, and he is without question the right person to lead Apple into the future. I could not be more confident in his abilities and his character, and I look forward to working closely with him on this transition and in my new role as executive chairman.

We reported just a couple of days ago that it strongly appears that the use of AI is becoming significantly more expensive.

Companies offering AI solutions have for years distributed the computing power required by AI to users almost for free, incurring huge losses themselves. Now it appears that all the biggest companies in the industry are tightening their policies and making users actually pay the true price for the AI resources used.

Now, the latest to join the ranks is Microsoft-owned GitHub. Its AI assistant, GitHub Copilot, was in many ways the first significant, modern AI product and was released long before ChatGPT.

GitHub stated that the explosive growth of so-called agentic programming has made its fixed monthly pricing model impossible (i.e., significantly unprofitable).

The company is immediately suspending the opening of all new GitHub Copilot accounts for individual users. The change affects all common price tiers, meaning GitHub Copilot Pro, Pro+, and Student are all frozen for new users for the time being.

Apple's upcoming Mac Studio and MacBook Pro models appear to be delayed by several months due to the global shortage of memory and storage components.

According to Bloomberg reporter Mark Gurman, who reported on the matter, the company is being forced to push back the launches of both new products, as the current lineup is starting to sell out.

For example, the current Mac Studios, which are sold in configurations equipped with M3 Ultra and M4 Max processors, are sold out in many places and delivery times are becoming very long, even if you order directly through Apple's online store. In addition, several configurations cannot be ordered at all at the moment. Previously, similar situations only occurred when the launch of the next device generation was just around the corner. According to Gurman, however, the information now suggests that the component shortage is simply limiting Apple's ability to manufacture and deliver devices.

According to Gurman, the next Mac Studio, which will likely be based on the M5-generation chips, was originally scheduled to appear in mid-summer 2026. Now the release is expected around October instead. The delay is explained in particular by the poor availability of RAM and SSDs, which has hit the entire technology industry hard. Mac Studio has become especially popular among users who run AI models locally and need significantly more memory and storage in their computers - precisely the components that are now in critically short supply.

WhatsApp is preparing a significant structural change for regular users. A new WhatsApp Plus subscription service has been revealed through the app's beta testing, and it is currently being tested with a limited group of users. It is an optional paid version that is heading towards an official release.

Although the feature is still in trial use, it reveals Meta's strategy: WhatsApp is following in the footsteps of its competitors, such as Telegram and Snapchat, by introducing a "freemium" model. Previously, Meta has introduced a paid version for Instagram and Facebook.

Based on the beta versions, WhatsApp Plus does not change the basic use of the app, but adds a layer on top of it that is aimed especially at power users and those who value personalization.

Key benefits include, among other things, the ability to pin up to 20 conversations to the top of the chat list instead of the current three. The test version has shown 18 new theme colors and 14 different app icons, with which the look of the app can be changed from green to, for example, purple or black and white. The package includes new ringtones and special stickers whose animations are displayed full-screen also for the message recipient.

In movie theaters, IMAX has been a kind of industry top standard. Dedicated moviegoers have known for decades that if you want to see a film on a massive screen, in all its glory, you should head to an IMAX theater.

But there is only a relatively small number of IMAX theaters worldwide.

And this, of course, creates the problem that if two blockbuster films are coming out, at the same time, that would benefit from an IMAX-level viewing experience, there's not enough IMAX theaters available.

And this is exactly what's happening this year. Disney would have wanted the Avengers: Doomsday film to be released in IMAX theaters, but the screens are already largely booked worldwide for Dune: Part Three, arriving in theaters on December 18.

Annoyed by the situation, Disney is establishing its own competitor for IMAX certification. The certification, going by the name Infinity Vision, would be granted to movie theaters that can offer "massive screens, with the sharpest and clearest colors and the best possible sound reproduction".

According to Kotaku, who reported on the matter, more detailed information about the new Infinity Vision certification has not yet been released. But Disney plans to get Infinity Vision certification for 75 movie theaters in the United States and 300 theaters worldwide.

Consumers have jumped on board as AI users worldwide, with clear evidence such as how ChatGPT quickly became the world's fastest-growing service.

Companies have also joined the bandwagon - especially those involved in software development. The end of 2025 is generally considered a major turning point in software development, when Claude Code made such a huge leap in AI-assisted programming that it managed to win over even those coders who had previously been skeptical of AI's programming skills.

But the hype has hidden an unpleasant truth that AI companies have barely dared to speak aloud.

Almost all current AI services are, in fact, loss-making businesses for the companies offering them. This means that consumers are currently able to use AI solutions that they might not be able to afford - or be willing to pay for.

Information about this financially unsustainable situation is hidden in companies' financial figures. And digging out precise data is made more difficult by the fact that the industry's largest independent players, OpenAI and Anthropic, are not publicly listed companies.

An essential part of business culture is that only a small portion of founded companies survive. The rest go bankrupt or their business just quietly shuts down.

Now these companies, whose business never really took off, have come up with a completely new trick to make a nice slice of money for their owners (or debtors).

Financial magazine Forbes reports (paywall) how defunct companies have started selling their former employees' email messages, Slack messages, and Teams messages as training material for AI companies' language models.

Apparently, AI companies are very interested in paying quite a nice sum of money for real, company-internal discussions. Presumably, this is because there is very little public material available from real working life that could be used for training language models.

According to Forbes, actors have already emerged in the market who help in shutting down a company - and at the same time also help sell the old company-internal messages to the highest-paying AI company. According to information, payments for company message collections range between $10,000 and $100,000, depending on the size of the message history and the number of employees.

Google has decided to bring badly behaving websites under better control and has announced a new change affecting visibility in its search engine.

Starting June 15, Google will automatically reduce the visibility in its search results of sites that "hijack" the user's browser back button for themselves.

The practice is quite common on even slightly dubious sites. It works so that when you try to go back from the site - for example to the search engine results - tapping the back button instead opens the site's front page, a prompt asking "are you really leaving our site", or perhaps an advertisement.

Google describes the problem in its own blog as follows:

When a user clicks the "back" button in the browser, they have a clear expectation: they want to return to the previous page. Back button hijacking breaks this fundamental expectation. It occurs when a site interferes with a user's browser navigation and prevents them from using their back button to immediately get back to the page they came from. Instead, users might be sent to pages they never visited before, be presented with unsolicited recommendations or ads, or are otherwise just prevented from normally browsing the web.

Anyone who has ever done a clean install of Windows on their computer knows that it is not a particularly smooth or fast-moving process.

But the update released for Windows 11 in April 2026 speeds up the installation of Windows significantly compared to before.

The trick is done simply by allowing the user, during the Windows installation, to choose that Windows updates will be installed only after the computer is already up and running.

Until now, the installation process has always required that the latest software and security updates are downloaded at the same time during installation - and installed as part of the Windows installation.

According to Windows Central, which reported on the matter (link in English), Microsoft nevertheless reminds users of the risks involved (link goes to X/Twitter): for example, security updates will not be up to date immediately when the computer is taken into use if installed this way.

Only this week we reported how the European Commission was touting its new age verification app as "technically ready" and top-class in terms of privacy. Now, however, the app has come under heavy criticism after a security researcher demonstrated that its protections are almost non-existent.

Security consultant Paul Moore says he bypassed the app's protections in under two minutes. According to him, the app's technical implementation contains elementary design flaws that make it an easy target for any technically savvy user.

Moore's analysis reveals that the app's trust model is based on local files that the user can edit themselves. Although the app is open source - something Commission President Ursula von der Leyen presented as a guarantee of transparency - this appears to have backfired by exposing its flimsy architecture.

According to Moore, the app stores the PIN code on the device, but it is not tied to the user's actual identity vault. By deleting certain values from the app's configuration file, an attacker can set a new PIN code and still gain access to the previous credentials.