The onrushing AI era was supposed to create boom times for great gadgets. Not long ago, analysts were predicting that Apple Intelligence would start a “supercycle” of smartphone upgrades, with tons of new AI features compelling people to buy them. Amazon and Google and others were explaining how their ecosystems of devices would make computing seamless, natural, and personal. Startups were flooding the market with ChatGPT-powered gadgets, so you’d never be out of touch. AI was going to make every gadget great, and every gadget was going to change to embrace the AI world.

This whole promise hinged on the idea that Siri, Alexa, Gemini, ChatGPT, and other chatbots had gotten so good, they’d change how we do everything. Typing and tapping would soon be passé, all replaced by multimodal, omnipresent AI helpers. You wouldn’t need to do things yourself; you’d just tell your assistant what you need, and it would tap into the whole world of apps and information to do it for you. Tech companies large and small have been betting on virtual assistants for more than a decade, to little avail. But this new generation of AI was going to change things.

There was just one problem with the whole theory: the tech still doesn’t work. Chatbots may be fun to talk to and an occasionally useful replacement for Google, but truly game-changing virtual assistants are nowhere close to ready. And without them, the gadget revolution we were promised has utterly failed to materialize.

In the meantime, the tech industry allowed itself to be so distracted by these shiny language models that it basically stopped trying to make otherwise good gadgets. Some companies have more or less stopped making new things altogether, waiting for AI to be good enough before it ships. Others have resorted to shipping more iterative, less interesting upgrades because they have run out of ideas other than “put AI in it.” That has made the post-ChatGPT product cycle bland and boring, in a moment that could otherwise have been incredibly exciting. AI isn’t good enough, and it’s dragging everything else down with it.

Archive link: https://archive.ph/spnT6

The Federal Trade Commission is moving to push back a trial set to determine if Amazon tricked customers into signing up for Prime subscriptions.

At a Zoom status hearing on Wednesday, the FTC officially asked US District Judge John Chun to delay the trial. According to the FTC's attorney, Jonathan Cohen, the agency needs two months to prepare beyond the September 22 start date, blaming recent "staffing and budgetary shortfalls" stemming from the Trump administration’s Department of Government Efficiency (DOGE), CNBC reported.

"We have lost employees in the agency, in our division, and on our case team," Cohen said, explaining that "there is an extremely severe resource shortfall in terms of money and personnel," Bloomberg reported. Cuts are apparently so bad, Cohen told Chun that the FTC is stuck with a $1 cap on any government credit card charges and "may not be able to purchase the transcript from Wednesday’s hearing," Bloomberg reported.

Further threatening to scramble the agency's trial preparation, the FTC anticipates that downsizing may require a move to another office "unexpectedly," Cohen told Chun.

Manufacturers are slowly starting to listen to what car journalists and owners have been complaining about for almost a decade: Cramming all the car’s functions into a touchscreen is an inferior solution to having dedicated physical controls for key tasks.

Among the manufacturers known to be switching back to buttons is Volkswagen, whose latest vehicles have gone touch-control-crazy with functions either buried inside a touchscreen menu or relocated to an annoying haptic feedback panel.

We’ve known for a while that Volkswagen was considering putting back some buttons in its cars, but the manufacturer never officially acknowledged this. Now VW’s design boss, Andreas Mindt, has admitted to Autocar that this approach was a mistake and that the automaker is backtracking on this trend.

“From the ID.2all onwards, we will have physical buttons for the five most important functions—the volume, the heating on each side of the car, the fans and the hazard light—below the screen,” Mindt told Autocar. He added, “They will be in every car that we make from now on. We will never, ever make this mistake anymore. On the steering wheel, we will have physical buttons. No guessing anymore. There's feedback, it's real, and people love this. Honestly, it's a car. It's not a phone.”

A survey of more than 2,000 smartphone users by second-hand smartphone marketplace SellCell found that 73% of iPhone users and a whopping 87% of Samsung Galaxy users felt that AI adds little to no value to their smartphone experience.

SellCell only surveyed users with an AI-enabled phone – thats an iPhone 15 Pro or newer or a Galaxy S22 or newer. The survey doesn’t give an exact sample size, but more than 1,000 iPhone users and more than 1,000 Galaxy users were involved.

Further findings show that most users of either platform would not pay for an AI subscription: 86.5% of iPhone users and 94.5% of Galaxy users would refuse to pay for continued access to AI features.

From the data listed so far, it seems that people just aren’t using AI. In the case of both iPhone and Galaxy users about two-fifths of those surveyed have tried AI features – 41.6% for iPhone and 46.9% for Galaxy.

So, that’s a majority of users not even bothering with AI in the first place and a general disinterest in AI features from the user base overall, despite both Apple and Samsung making such a big deal out of AI.

A contractor for Immigration and Customs Enforcement (ICE) and many other U.S. government agencies has developed a tool that lets analysts more easily pull a target individual’s publicly available data from a wide array of sites, social networks, apps, and services across the web at once, including Bluesky, OnlyFans, and various Meta platforms, according to a leaked list of the sites obtained by 404 Media. In all the list names more than 200 sites that the contractor, called ShadowDragon, pulls data from and makes available to its government clients, allowing them to map out a person’s activity, movements, and relationships.

ShadowDragon says in marketing material its tools can be used to monitor protests, and claims it found protests around Union Station in Washington DC during a 2023 visit by Benjamin Netanyahu. Daniel Clemens, ShadowDragon’s CEO, previously said on a podcast that protesters should not “be surprised when people are going to investigate you because you made their life difficult.”

“The long list of sites and services that ShadowDragon’s SocialNet tool accesses is a reminder of just how much data is accessible and collected from and about us to provide surveillance services to the government and others,” Jeramie Scott, senior counsel and director the Electronic Privacy Information Center’s (EPIC) Project on Surveillance Oversight, told 404 Media in an email. “SocialNet is just one example of the unchecked surveillance ecosystem that lacks any meaningful transparency, oversight, or accountability that allows the government to circumvent Constitutional and statutory protections to access sensitive personal data,” he added.

The leaked list of targeted sites and services include ones from major tech companies such as Apple, Amazon, Meta, Microsoft, and TikTok. It also includes communication tools like Discord and WhatsApp; activity- or hobby-focused sites like AllTrails, BookCrossing, Chess.com, and cigar review site Cigar Dojo; payment services like Cash App, BuyMeACoffee, and PayPal; sex worker sites OnlyFans and JustForFans; and social networks Bluesky and Telegram. Even relatively obscure social networks are included in the list, such as BeReal.

A contractor for Immigration and Customs Enforcement (ICE) and many other U.S. government agencies has developed a tool that lets analysts more easily pull a target individual’s publicly available data from a wide array of sites, social networks, apps, and services across the web at once, including Bluesky, OnlyFans, and various Meta platforms, according to a leaked list of the sites obtained by 404 Media. In all the list names more than 200 sites that the contractor, called ShadowDragon, pulls data from and makes available to its government clients, allowing them to map out a person’s activity, movements, and relationships.

ShadowDragon says in marketing material its tools can be used to monitor protests, and claims it found protests around Union Station in Washington DC during a 2023 visit by Benjamin Netanyahu. Daniel Clemens, ShadowDragon’s CEO, previously said on a podcast that protesters should not “be surprised when people are going to investigate you because you made their life difficult.”

“The long list of sites and services that ShadowDragon’s SocialNet tool accesses is a reminder of just how much data is accessible and collected from and about us to provide surveillance services to the government and others,” Jeramie Scott, senior counsel and director the Electronic Privacy Information Center’s (EPIC) Project on Surveillance Oversight, told 404 Media in an email. “SocialNet is just one example of the unchecked surveillance ecosystem that lacks any meaningful transparency, oversight, or accountability that allows the government to circumvent Constitutional and statutory protections to access sensitive personal data,” he added.

The leaked list of targeted sites and services include ones from major tech companies such as Apple, Amazon, Meta, Microsoft, and TikTok. It also includes communication tools like Discord and WhatsApp; activity- or hobby-focused sites like AllTrails, BookCrossing, Chess.com, and cigar review site Cigar Dojo; payment services like Cash App, BuyMeACoffee, and PayPal; sex worker sites OnlyFans and JustForFans; and social networks Bluesky and Telegram. Even relatively obscure social networks are included in the list, such as BeReal.

Google has become so integral to online navigation that its name became a verb, meaning "to find things on the Internet." Soon, Google might just tell you what's on the Internet instead of showing you. The company has announced an expansion of its AI search features, powered by Gemini 2.0. Everyone will soon see more AI Overviews at the top of the results page, but Google is also testing a more substantial change in the form of AI Mode. This version of Google won't show you the 10 blue links at all—Gemini completely takes over the results in AI Mode.

This marks the debut of Gemini 2.0 in Google search. Google announced the first Gemini 2.0 models in December 2024, beginning with the streamlined Gemini 2.0 Flash. The heavier versions of Gemini 2.0 are still in testing, but Google says it has tuned AI Overviews with this model to offer help with harder questions in the areas of math, coding, and multimodal queries.

With this update, you will begin seeing AI Overviews on more results pages, and minors with Google accounts will see AI results for the first time. In fact, even logged out users will see AI Overviews soon. This is a big change, but it's only the start of Google's plans for AI search.

Gemini 2.0 also powers the new AI Mode for search. It's launching as an opt-in feature via Google's Search Labs, offering a totally new alternative to search as we know it. This custom version of the Gemini large language model (LLM) skips the standard web links that have been part of every Google search thus far.

It was a strong year for Best Animated Feature Oscar nominations, and an underdog triumphed. At the 97th annual Academy Awards at Ovation Hollywood in Los Angeles last night, Flow beat competition from Pixar's Inside Out 2, DreamWorks' The Wild Robot and Aardman's Wallace & Gromit: Vengeance Most Fowl.

Gints Zilbalodis tale about a cat in a flooded world missed out on the Oscar for Best International Feature Film but still became Latvia's first Oscar win. And it was surely also the first Oscar winner to be made entirely in the free 3D modelling software Blender, cementing the open-source program's place among the best animation software.

Flow was one of our highlights of Annecy 2024, and it still seems incredible that it was made by a small team using Blender alone. It was rendered in EEVEE, Blender's realtime render engine.

Gints thanked Blender when accepting the award. Speaking to press afterwards, he said: "Any kid now has tools that are used to make now Academy Award-winning films, so I think we're going to see all kinds of exciting films being made from kids who might not have had a chance to do this before.

A team of volunteer archivists has recreated the Centers for Disease Control website exactly as it was the day Donald Trump was inaugurated. The site, called RestoredCDC.org, went live Tuesday and is currently being hosted in Europe.

As we have been following since the beginning of Trump’s second term, websites across the entire federal government have been altered and taken offline under this administration’s war on science, health, and diversity, equity, and inclusion. Critical information promoting vaccines, HIV care, reproductive health options including abortion, and trans and gender confirmation healthcare have been purged from the CDC’s live website under Trump. Disease surveillance data about bird flu and other concerns have either been delayed or have stopped being updated entirely. Some deleted pages across the government have at least temporarily been restored thanks to a court order, but the Trump administration has added a note rejecting “gender ideology” to some of them.

“Our goal is to provide a resource that includes the information and data previously available,” the team wrote. “We are committed to providing the previously available webpages and data, from before the potential tampering occurred. Our approach is to be as transparent as possible about our process. We plan to gather archival data and then remove CDC logos and branding, using GitHub to host our code to create the site.”

Apple reportedly filed an appeal in hopes of overturning a secret UK order requiring it to create a backdoor for government security officials to access encrypted data.

"The iPhone maker has made its appeal to the Investigatory Powers Tribunal, an independent judicial body that examines complaints against the UK security services, according to people familiar with the matter," the Financial Times reported today. The case "is believed to be the first time that provisions in the 2016 Investigatory Powers Act allowing UK authorities to break encryption have been tested before the court," the article said.

Although it wasn't previously reported, Apple's appeal was filed last month at about the time it withdrew ADP from the UK, the Financial Times wrote today.

"The case could be heard as soon as this month, although it is unclear whether there will be any public disclosure of the hearing," the FT wrote. "The government is likely to argue the case should be restricted on national security grounds."

At launch, access to Mullvad Leta was restricted to users with a paid Mullvad VPN account, but it is now free and open to all.

Mullvad Leta has been audited by Assured.

• FAQ
• Terms of Service

Just a heads up, some of the details in the FAQ and Terms of Service seem a bit outdated and might not be accurate anymore.

Some relevant information from their FAQ section is as follows:

What can I do with Leta?

Leta is a search engine. You can use it to return search results from many locations. We provide text search results, currently we do not offer image, news or any other types of search result. Leta acts as a proxy to Google and Brave search results. You can select which backend search engine you wish to use from the homepage of Leta.

Can I use Leta as my default search engine?

Yes, so long as your browser supports changing default search engines.

Navigate to https://leta.mullvad.net/ in your browser and right-click on the URL bar.

From there you should see Add “Mullvad Leta“ with the Mullvad VPN logo to the left.

If you do not see this, you can attempt to add a custom search engine to your browser with:

• The name set to: Leta
• The URL set to: https://leta.mullvad.net/?q=%25s

You can select which backend engine to use as follows:

• Google: https://leta.mullvad.net/?q=%25s&engine=google
• Brave: https://leta.mullvad.net/?q=%25s&engine=brave

Did you make your own search engine from scratch?

We did not, we made a front end to the Google and Brave Search APIs.

Our search engine performs the searches on behalf of our users. This means that rather than using Google or Brave Search directly, our Leta server makes the requests.

Searching by proxy in other words.

What is the point of Leta?

Leta aims to present a reliable and trustworthy way of searching privately on the internet.

However, Leta is useless as a service if you use the perfect non-logging VPN, a privacy focussed DNS service, a web browser that resists fingerprinting, and correlation attacks from global actors. Leta is also useless if your browser blocks all cookies, tracking pixels and other tracking technologies.

For most people Leta can be useful, as the above conditions cannot ever truly be met by systems that are available today.

What is a cached search?

We store every search in a RAM based cache storage (Redis), which is removed after it reaches over 30 days in age.

Cached searches are fetched from this storage, which means we return a result that can be from 0 to 30 days old. It may be the case that no other user has searched for something during the time that you search, which means you would be shown a stale result.

What happens to everything I search for?

Your searches are performed by proxy, it is the Leta server that makes calls to the Google or Brave Search API.

Each search that has not already been cached is saved in RAM for 30 days. The idea is that the more searches performed, the larger and more substantial the cached results become, therefore aiding with privacy.

All searches will be stored hashed with a secret in a cache. When you perform a search the cache will be checked first, before determining whether a direct call to Google or Brave Search should be made. Each time the Leta application is restarted (due to an upgrade, or new version) server side, a new secret hash is generated, meaning that all previous search queries are no longer visible to Leta

What could potentially be a unique search would become something that many other users would also search for.

What is running on the server side?

We run the Leta servers on STBooted RAM only servers, the same as our VPN servers. These servers run the latest Ubuntu LTS, with our own stripped down custom Mullvad VPN kernel which we tune in-house to remove anything unnecessary for the running system.

The cached search results are stored in an in-memory Redis key / value store.

The Leta service is a NodeJS based application that proxies requests to Google or Brave Search, or returns them from cache.

We gather metrics relating to the number of cached searches, vs direct searches, solely to understand the value of our service.

Additionally we gather information about CPU usage, RAM usage and other such information to keep the service running smoothly.

At launch, access to Mullvad Leta was restricted to users with a paid Mullvad VPN account, but it is now free and open to all.

Mullvad Leta has been audited by Assured.

• FAQ
• Terms of Service

Just a heads up, some of the details in the FAQ and Terms of Service seem a bit outdated and might not be accurate anymore.

Some relevant information from their FAQ section is as follows:

What can I do with Leta?

Leta is a search engine. You can use it to return search results from many locations. We provide text search results, currently we do not offer image, news or any other types of search result. Leta acts as a proxy to Google and Brave search results. You can select which backend search engine you wish to use from the homepage of Leta.

Can I use Leta as my default search engine?

Yes, so long as your browser supports changing default search engines.

Navigate to https://leta.mullvad.net/ in your browser and right-click on the URL bar.

From there you should see Add “Mullvad Leta“ with the Mullvad VPN logo to the left.

If you do not see this, you can attempt to add a custom search engine to your browser with:

• The name set to: Leta
• The URL set to: https://leta.mullvad.net/?q=%25s

You can select which backend engine to use as follows:

• Google: https://leta.mullvad.net/?q=%25s&engine=google
• Brave: https://leta.mullvad.net/?q=%25s&engine=brave

Did you make your own search engine from scratch?

We did not, we made a front end to the Google and Brave Search APIs.

Our search engine performs the searches on behalf of our users. This means that rather than using Google or Brave Search directly, our Leta server makes the requests.

Searching by proxy in other words.

What is the point of Leta?

Leta aims to present a reliable and trustworthy way of searching privately on the internet.

However, Leta is useless as a service if you use the perfect non-logging VPN, a privacy focussed DNS service, a web browser that resists fingerprinting, and correlation attacks from global actors. Leta is also useless if your browser blocks all cookies, tracking pixels and other tracking technologies.

For most people Leta can be useful, as the above conditions cannot ever truly be met by systems that are available today.

What is a cached search?

We store every search in a RAM based cache storage (Redis), which is removed after it reaches over 30 days in age.

Cached searches are fetched from this storage, which means we return a result that can be from 0 to 30 days old. It may be the case that no other user has searched for something during the time that you search, which means you would be shown a stale result.

What happens to everything I search for?

Your searches are performed by proxy, it is the Leta server that makes calls to the Google or Brave Search API.

Each search that has not already been cached is saved in RAM for 30 days. The idea is that the more searches performed, the larger and more substantial the cached results become, therefore aiding with privacy.

All searches will be stored hashed with a secret in a cache. When you perform a search the cache will be checked first, before determining whether a direct call to Google or Brave Search should be made. Each time the Leta application is restarted (due to an upgrade, or new version) server side, a new secret hash is generated, meaning that all previous search queries are no longer visible to Leta

What could potentially be a unique search would become something that many other users would also search for.

What is running on the server side?

We run the Leta servers on STBooted RAM only servers, the same as our VPN servers. These servers run the latest Ubuntu LTS, with our own stripped down custom Mullvad VPN kernel which we tune in-house to remove anything unnecessary for the running system.

The cached search results are stored in an in-memory Redis key / value store.

The Leta service is a NodeJS based application that proxies requests to Google or Brave Search, or returns them from cache.

We gather metrics relating to the number of cached searches, vs direct searches, solely to understand the value of our service.

Additionally we gather information about CPU usage, RAM usage and other such information to keep the service running smoothly.