Friday, June 14th 2024
Microsoft Delays Controversial "Recall" Feature for Windows 11 24H2
Microsoft has made a last-minute decision to pull its much-debated "Recall" feature from the Windows 11 24H2 update set to launch on June 18th. Instead, the company will roll out Recall as a preview through the Windows Insider Program while it works to build user trust and address security concerns. Recall, one of the flagship features of 24H2, creates a searchable 30-day timeline of a user's activities including files, webpages, and screenshots. However, since its announcement on May 20th, Recall has faced heavy criticism over potential privacy risks from storing user data in unencrypted plain text files. Security researcher Kevin Beaumont labeled Recall a "security nightmare" after finding it logged activities to a SQLite database accessible by non-admin accounts. This raised alarms about the depths of user behavior tracked and stored locally on PCs.
Initially, Microsoft had planned for Recall to be enabled by default in 24H2. However, following the backlash, the company backtracked on June 7th, making it an opt-in feature requiring Windows Hello authentication and adding encryption. Those adjustments were still not enough to satisfy Microsoft. In a new blog post, the firm stated Recall did not yet meet its "own standards of quality and security" and that it "must be trustworthy, secure and robust" before a wider rollout. By moving Recall to the Insider Program for further testing and refinement, Microsoft is giving itself more time to get the technology right and rebuild user confidence. A future blog will provide instructions for Insiders to preview Recall on compatible Copilot+ PCs with added security protections.
Source:
Microsoft
Initially, Microsoft had planned for Recall to be enabled by default in 24H2. However, following the backlash, the company backtracked on June 7th, making it an opt-in feature requiring Windows Hello authentication and adding encryption. Those adjustments were still not enough to satisfy Microsoft. In a new blog post, the firm stated Recall did not yet meet its "own standards of quality and security" and that it "must be trustworthy, secure and robust" before a wider rollout. By moving Recall to the Insider Program for further testing and refinement, Microsoft is giving itself more time to get the technology right and rebuild user confidence. A future blog will provide instructions for Insiders to preview Recall on compatible Copilot+ PCs with added security protections.
81 Comments on Microsoft Delays Controversial "Recall" Feature for Windows 11 24H2
It's not just an engineering failure. It's not just a design failure, PR failure, or marketing failure. It wasn't a timing error or some sort of demo glitch.
This is a high level leadership failure. All of Microsoft's senior executives reviewed the Recall feature before it was unveiled. Unless they are in some of purely administrative role, all of Microsoft's SVP level executives should be ashamed this happened. This includes General Counsel and everyone in the C-level.
It wasn't just a minor oversight or some nit-picky thing that only a security researcher could have identified. There are multiple problems with the conception and design of Recall. Basic issues that apparently no one at Microsoft was aware of and no one had any legitimate responses for in the first few days.
They should have anticipated all of the basic questions that came up particularly about privacy. They should have had answers queued up. They did nothing. They unveiled some half-baked teenager project (which discredits a lot of more thoughtful high schoolers) and couldn't reply to any of the simplest security concerns.
Why would anyone trust Microsoft to do things correctly under the hood anywhere else in their software/service stack?
If you are CTO at some big financial institution, pharmaceutical company, manufacturer, government agency, whatever, do you believe Microsoft understands privacy? Security? Software QA?
Regardless of where you live, as a taxpayer do you want your government's tax agency to run their systems on Microsoft Windows?
At this point, one should really ask the Microsoft Board of Directors, "do you have the right team running this company?" Start with the CEO which stands for Chief Executive Officer. How competently is that person executing right now?
Here's my guess. Microsoft will rewrite Recall, release it right before Christmas 2024 and it will have less than half the functionality shown in the original announcement. It will do far less, run poorly and eventually get dropped from Windows because no one will find particularly useful -- just like Cortana. It'll end up in the big graveyard of Failed Microsoft Consumer Technologies, buried under a pile of Zune music players and Windows Mobile phones.
If you think of a PC as a box that you can plug a CPU, memory, drives, and graphics card into, you are part of a very small subset of users. Over 85% of Apple computers are notebook models. And here in 2024 most comparable Windows notebooks are built similarly to MacBooks: with no serviceable parts.
Customer satisfaction surveys are all self reported and Apple computer users have much higher customer satisfaction scores than Windows users. Remember that there are many people who use both (I'm one). In fact, most Mac households have a Windows PC (work or personal) under the same roof.
From an end user satisfaction standpoint, Linux is way worse of a desktop operating system than both. Linux is appallingly bad on the desktop (and even worse in notebook computers due to piss poor power management) and I installed Linux the first time way back in 1998.
Anyhow, this discussion is becoming more moot with each passing week. The primary computing modality for consumers in 2024 is the smartphone not PC. Microsoft has zero presence in this area now because they stupidly abandoned Windows Mobile.
Smartphones are the driver of consumer electronics innovation, both the software and hardware side. It is going to be the same for AI adoption. Consumer AI innovation will be driven by smartphones: performance-per-watt, memory requirements, latency, model complexity, integration with a constantly changing personal relevance.
And this is absolute horror, for the companies which work with secret, or private projects, and have to use the Windows. The amount of private data, the project NDA data, etc, which could potentialy "leak" regardless the encryption, just because MS has a habit, to let the bugs "accidetnally""slip" into their updates. Of course, the companies most likely rely on Enterprise editions, including the employee's PCs. But what if someone works for oneself, and the entire labour can be just watched over and scanned real time, and sent to the HQ. Simultaneously feeding the LLM, with the data, that is didn't get the copyright protection yet. I'm afraid, not for long. They'll find the way to make it some hidden task, not even showing in process list, or like recently, a part of other processes, like svhost, and other crucial core executables, making it mandatory to run. As much as adding new servers in the host, with each windows update, so users cannot avoid their Windows OS to constantly "calling home".
So, in other words, the foundation has been set many years ago. This is true. However Ai is sadly here to stay. The users being forced, regardless of their free will. The more people will stay in the MS ecosystem, the more they will wind themselves in the quicksands. The other problem is, the moment MS Windows (theoretically) popularity would slip from them, the AI stuff with switch to whatever next popular OS.
Unfortunately, but cvaldes has a point here. No matter what people want, as much as people use the Windows OS, MS is able to do, whatever they want, and even stop supporting the HW, they don't like, so all current ones might render moot, and using it might indeed be akin to dreaming.
The problem is, many Windows users, are as much hostages, as Apple users. Many are just locked int their ecosystem, many just willingly support MS and their business, no matter what. Apple just gives it's users a fancy premium illusion of freedom, a posh gaslighting, with silver lining. While Microsoft, just does this boldly and openly. Both companies have became the most valuable entities, not because of their great products, or services. The money lie into data mining and resell.
We know from consumer satisfaction surveys that Apple stands above the competition, both on the computing side as well as the smartphone side.
While I've never owned an Android phone, I've certainly used other non-Apple handsets over the years. And I've certainly used a bunch of computers with a variety of operating systems including every single version of macOS/OS X, every single consumer Windows, a bunch of Linux distros, FreeBSD, a couple of commercial Unixes, Apple System 6 & System 7, even some older operating systems.
And while the nature of Apple's consumer business has definitely evolved over the years, I still have the option of taking my Mac/iPhone/iPad to an nearby Apple Store and have someone help troubleshoot a problem. How are those Microsoft Stores doing? Any nearby in 2024?
And like I mentioned earlier, there are Macs and Windows PCs in my house. I paid for all of them. And yet here I am typing this reply on my Mac...
Apple vertically integrated product stack (hardware, software, services) will benefit massively from systemwide AI integration. After a couple of years of trepidation and hesitation (which is a very sensible reaction), many people will realize the benefit of AI assistance on a large scale.
It's not about the computer on your desk. It's about the phone in your pocket, the watch on your wrist, the earbud, and maybe other things (some of which are unreleased and sitting in a lab in Cupertino).
Microsoft has no presence whatsoever on a person's body when they step out of the house.
Like I said earlier, consumer AI innovation will be driven by the smartphone. What can be done on a small battery powered device with limited memory? That's where the consumer AI battle will be duked out, not on $2500 GPUs running off of 1200W PSUs in some nerd cave.
If you want to program AI apps for iOS/iPadOS devices, yes, by all means get an M4 Mac.
But if you just want AI to help you with your daily life as a consumer and maybe do some work related stuff, you're probably better off waiting for the next gen iPhone 16 when it releases this fall.
Personally I plan on waiting for the M5 SoC chips before I upgrade my Mac mini M2 Pro. But my iPhone will be replaced first. That's the first device I grab before I step out the front door. That's the one that truly matters.
Microsoft completely screwed their future consumer prospects when they canceled Windows Phone.
EDIT: I would not say that, they just like the taste of money. It's not been what the people who use the OS for so many years now and why should it be as they are making a killing without making it for the people who use it.
It's about what they can get away with not what the people want.
What a horrible idea from none other than Microsoft I agree, encryption is only good till it’s broken. There is a reason so much encrypted data is saved by our government and it’s merely to break the encryption and then blackmail and control people.
I will point out that most businesses, especially smaller ones live on a razor thin margins. This includes pretty much all restaurants and most service industry businesses. A few percentage points may end up being enough to keep those businesses afloat. While there are no AI tools for that mom-and-pop restaurant right now, they will come soon enough.
If you have ever worked for a small business and understood the business/financial side of things, you should know that margins are very, very tight. Even adding one extra employee might tip the costs into the red (salary, healthcare, benefits, etc.). That's the fine line many small business owners are walking on.
Quick, what's the number one annoyance of a restaurant manager? Scheduling employee shifts. Person A has school these nights, Person B has to pick up Kid 1 at this time, Kid 2 at this time. Person C has another job on these days. Person D is on call. Then sick days, vacations, seasonal demand. If the AI scheduler knows there's a concert one night, shouldn't it remind the operator that they might need a couple of people on deck? Working back to back shifts? If someone closes late one night, try not to schedule them to open the following morning. People still may want prime shifts (Friday and Saturday night, Thursday happy hour, Sunday brunch).
Someday in the near future, a lot of the work scheduling will be done by AI tools. And not just in a restaurant. Right now it's a human. They pore over various requests/requirements, post a schedule and hope that no one brings up an issue that requires a serious revision.
That's just one example in one industry. Again, a lot of the silly consumer AI frivolities will die off quickly. But businesses will start to see how to use this effectively. And if AI helps you write a business e-mail more effectively at some big company, the little guy will want it as well. And if your AI-assisted phone can make it sound like you have an executive assistant who has been there for 25 years, sure, why not?
For those who think AI is a passing fad, it's not. Fortune 500 companies are already using this. It's not just Google using it for auto image captioning (they started doing that 10 years ago).
I don't think AI is a passing fad, but I don't think it's beneficial in the final analysis, either. In my list of reasons why, I didn't even mention the deluge of cheating/scamming/spoofing that it enables, in schools and elsewhere. Sometimes I almost wish AI really were an apocalyptic singularity, an impossibly smart and potentially hostile human-style consciousness. That would not only be far more impressive than what we have now; at least the threat of it would be serious and straightforward, rather than the death-by-a-thousand-frivolous-cuts that we seem to be heading towards.
But hey, business growth line goes up, so let's all celebrate.
However the datamining example is good. If some business analyst in a team of five retires or moves on, maybe some of his workload can be taken on by an employee doing split work (half in the analyst role, half in something else) by using AI.
AI isn't going to instantly make people disappear from an office or production line. It will give the chance to put high value people doing things that are more valuable. In my restaurant example, it would be better if said restaurant manager spend less time writing the upcoming weekly schedule and spend more time training the next host/greeter on the restaurant floor or maybe doing a walkthrough with a good corporate customer's event planner or having a quick drink with them.
It's really how astutely the technology is used. One business might install AI tools and promptly do all the wrong things with them. Another business might be more cautious and thoughtful about what tasks to assign to the AI and what to leave to experienced employees.
It's really the same with any technological advancement. There were a lot of doomsayers that thought the Industrial Revolution would put everyone out of work. As we know, that was not the case. Certainly some roles and positions have vanished but others have taken their place. Not everyone needs to shovel coal into a boiler, right?
Even a more recent technological innovation like social media which some erroneously dismiss as worthless. There are plenty of examples of bad usage, but also some wonderful examples of good usage. Even mundane stuff like your city's police department tweeting, "Avoid the intersection of 2nd and Main, there has been an injury accident." That's something that helps everyone out. You avoid traffic, first responders have less distractions, etc.
But for sure, AI is not going away. It's up to humans to use their best judgment how to use the technology. Some people will use it for nefarious purposes. Others will get it quickly and do the right thing. A lot of people will wander between using it well and using it not so well.
They needed to unveil a well thought out AI experience that addressed basic and valid concerns about privacy and security. They failed.
Nope, Microsoft went for the glitzy demo and then had nothing to say for a few days when people kicked the tires.
As I said earlier, this is not Microsoft's first rodeo. They aren't some little startup just tryin' to figure stuff out because y'all know this computing stuff can be a little tricky at times.
For sure Microsoft will try to ramrod Recall down everyone's throats just like they tried with Cortana. My guess is when it debuts, it will do so rather weakly, a few people will play around with it for a month or two, then the world will move onto something else because Microsoft will have very limited AI presence on their phones. Out of sight, out of mind. "Microsoft? Who cares? It doesn't run on my phone."
Two years from now, Microsoft will rebrand it, try to relaunch it, and it will fail again. They'll let it fester for a couple more years then pull the plug.
Because that's the Microsoft Way.
:):p:D
>>webpages, and screenshots....
A decision was made recently to move from Windows to Linux Ubuntu as a primary software development platform. Transition to Linux Ubuntu will take some time, of course, however No way to accept such monitoring of activities.
Just screenshots is next to useless for users, great for big brother, hr and other of the ilk