Shoutout to the marketing ‘expert’ who mailed a large, battery-powered, LCD-equipped video greeting card to our CEO to demonstrate their innovative approach to ignoring the #ewaste crisis. Bold strategy.
We’ll be sure to reach out if we need help shipping steaks and leather handbags to PETA.
A bunch of tech CEOs are talking about being “AI-first”, the latest in a series of CEO fads like “return to office”, “founder mode” and “pivot to full fascism”. This one’s weird, though, because it only makes sense if… none of their workers are great at their jobs. anildash.com//2025/04/19/ai-fi…
Saying that 30% of all your code is generated “by software” (not “by AI”, but “by software”) is what you’d say if you knew perfectly well you can’t write and maintain a production OS with LLMs, but you have a financial motivation to inflate the number to ride a hype wave and eyeballed how much of your committed code is machine-generated by templates techcrunch.com/2025/04/29/micr…
EDIT: it seems this reporting omitted the rather load-bearing phrase "in some of our projects" from his exact words:
> I’d say maybe 20 to 30 percent of the code that is inside of our repos today in some of our projects are probably all written by software.
Now, can I believe that "some" projects at Microsoft are 30% AI-generated? Yes, because they have a lot of projects at a lot of different levels of needs-to-workness.
Next in Penpot Desktop: one more step towards a seamless offline experience. With a click of a magic wand, you will be able to set up a local @penpot instance.
It requires Docker and will launch as an experimental feature.
“The life cycle has completed itself: Video killed the radio star. The radio star became a podcaster, the podcaster launched a YouTube channel, and YouTube became television. Now streaming platforms have become cable bundles again”
That server is hosted in a hackerspace, and we _should_ get a new internet connection today. So, you're probably noticing that ;)
Will look into it later tonight; hopefully it'll be online late this evening or tomorrow the latest. Hold on to your issue, or send an email to info@mapcomplete.org
"Currently, in our bank account, we have £2100 (around $2800). It's not even enough to pay the rent for one month," explains Joseph Redon, head of the Japanese Game Preservation Society. He's describing the end of the non-profit organisation he set up in 2011 with the aim of properly safeguarding Japanese gaming history before it is lost forever.
"I'm paying 25% of the rent," reveals Redon. "Meaning we are already virtually out of business. We have monthly supporters but are losing between £260 and £420 every month. So when you do the maths, we are completely out of cash in September. And a non-profit organization cannot be out of cash. We're then out of business. What we really need is to bring in at least £750 every month. It's not that much if it's 400 people giving £2 a month. Some people might give more, so we need 300 people to become supporters so that we can stay afloat. Otherwise, we shut the laboratory and won't have a place to do preservation or archiving."
Overview
This allows the user to select which algorithms are presented in the signature_algorithms client hello extension.
In this change, I add the CURLOPT_SSL_SIGNATURE_ALGORITHM option for curl_...
Taky to ten Garmin může hlásit dřív, že je vybitá baterie v pedálech, ještě když je tam baterie 🔋 CR1/3N, kterou samozřejmě nikde nemají. Tak ji musím objednat.
Nebudeš tomu věřit, ale měnil jsem ji v pedálu před pár dny. Měl jsem jednu rezervní, ale raději jsem ji hned objednal do zásoby. Mám zkušenost, že mi po první signalizaci úplně umřela v půlce třetího švihu.
Já objednal taky raději dvě. Ani mě nenapadlo, že ji budu muset objednávat, pak na ní koukám a on to takový divný váleček 😃 No snad dorazí do pátku, ale zítra je svátek, tak moc tomu nevěřím. Ještě v pátek dávám kolo do servisu.
French people, seeing the Canadians happy that the insane candidate was not elected, but troubled because said insane candidate still got 40% of the vote:
Since 1.10.2 the BBL network plugin ships with debugger detecting code that will terminate the application when a debugger is attached. This makes it impossible to run the slicer from Visual Studio...
While I like Tuta very much I've to say that the E2E encryption is only possible via pre-shared passwords or when sender and recipient both are using Tuta.
Tuta is not supporting established E2E crypto as e.g. GPG. This results in a limited applicability of E2E encryption and IMHO results in most mails send from Tuta users are not E2E encrypted but only use transport crypto (TLS).
@voidstar Thanks for your feedback! You're right — full E2E encryption with Tuta requires either both users on Tuta or a shared password for external recipients. We chose built-in encryption over PGP to ensure privacy is automatic and user-friendly. While TLS is used with non-Tuta users, we encourage switching to Tuta for full E2E by default.
Yes, I understand your rational but I think it's flawed.
I agree that some users cannot use GPG due it's complexity, but your approach is just not feasible. Pre-sharing a password with attorneys, doctors, authorities, etc does not work. So while in theory your approach might make sense it does not work leading to the results described.
Moreover, I think it should be the customers choice to use IMAP/POP with GPG if one wants to as both approaches are not mutually exclusive.
not looking for free email, but I can not pay via banking card. The payment method should either be somehow via cash or if not possible then via Monero.
Trotz Überschrift ist der Kommentar deutschlandfunk.de/kommentar-z… vom #DLF (leider nur auf seltsamen social media unterwegs) zur elektronischen Patientenakte auf weiten Strecken gut und ausgewogen.
Aber dann der Schluss: Man müsse halt auch mal mit halbgaren Lösungen starten, sonst bliebe nur das Faxgerät … 🤯
Ob die Kommentatorin in ein Auto steigt, bei dem man sich halt noch nicht um die Bremsen gekümmert hat? Und in einer Wohnung wohnt, bei dem das Dach noch nicht gebaut wurde?
Die elektronische Patientenakte startet. Trotz Datenschutzsorgen und technischer Hürden bleibt sie ein überfälliger Schritt für ein modernes Gesundheitswesen.
"Privacy on Trial: Meta’s DOJ Battle vs. Purism’s User-Centric Philosophy"
Meta claims it’s not a monopoly.
The FTC disagrees.
At the center of the case: Meta’s acquisitions of Instagram and WhatsApp—moves the FTC argues were calculated to "neutralize competitive threats" and consolidate power across the digital landscape.
For those who reject Surveillance Capitalism—where your data is the product—Purism offers a bold alternative.
Purism makes premium phones, laptops, mini PCs and servers running free software on PureOS. Purism products respect people's privacy and freedom while protecting their security.
Hnutí ANO by spolu s SPD, Stačilo! a Motoristy dosáhlo ve Sněmovně na ústavní většinu 122 poslanců, ukázal průzkum STEM pro CNN Prima News zveřejněný v neděli.
The Accessibility Task Force of the W3C Publishing Community Group has announced the release of the Accessibility Metadata Display Guide for Digital Publications 2.0. […]
At Blind Ambition, we are always excited to see new technology that has the potential to transform everyday life for blind and partially sighted people. The recent updates to the Ray-Ban Meta smart glasses promise to be a real game-changer.
The author of this piece from Blind Ambition states:
“It’s important to remember that, as empowering as this technology is, it should ideally be used alongside traditional mobility tools, such as a long cane, for maximum independence and safety.”
To use the word ideally is a bit of an understatement. It is absolutely essential that blind people who use these smart devices while navigating must continue to use their primary mobility aid at all times, whether we’re talking about a white cane or a guide dog. Smart glasses cannot and will never replace these critical mobility solutions. I cannot emphasize this enough. Also, as magical as these devices may seem to be, remember that AI still has a propensity to hallucinate. That’s a very kind way of saying that it can and will lie to you. As a blind person who regularly uses AI tools, I am well aware of how this technology has benefited our community but it’s important to always be aware of the software’s limitations.
At Blind Ambition, we are always excited to see new technology that has the potential to transform everyday life for blind and partially sighted people. The recent updates to the Ray-Ban Meta smart glasses promise to be a real game-changer.
@lulu_bear Blind Ambition responded very quickly to my feedback and have since updated their post to provide a more balanced and accurate perspective. I'm grateful to them for taking my feedback seriously and for responding so promptly. Here is the relevant portion of their blog post which they have revised.
It’s important to remember that, as empowering as this technology is, it must be used alongside traditional mobility aids, such as a long cane or a guide dog. This is not just a recommendation — it is absolutely essential. Smart glasses, no matter how advanced, cannot and will never replace critical mobility tools that have been developed over decades to ensure safety and independence for blind and partially sighted people. We cannot emphasise this enough.
Moreover, as magical as these AI-powered devices may seem, it’s important to acknowledge their limitations. Artificial Intelligence has a well-documented tendency to hallucinate — a polite way of saying that it can and does present inaccurate or entirely false information. As a blind person who regularly uses AI tools, I’m fully aware of the ways this technology benefits our community, but it’s vital to stay aware and cautious of these software limitations. Always verify information using multiple tools or human support where possible.
At Blind Ambition, we are always excited to see new technology that has the potential to transform everyday life for blind and partially sighted people. The recent updates to the Ray-Ban Meta smart glasses promise to be a real game-changer.
Latest maintenance release brings improved stability and fixes to the powerful free office suite Berlin, 30 April 2025 – The Document Foundation is pleased to announce the release of LibreOffice 25.2.
Love the alt text, but it doesn't actually tell you any of the figures. Felt a bit like clickbate in alt text form, which is incredibly funny as an idea if nothing else.
FSCast #257. Taming JAWS On the Web, an introduction to Innosearch, and more stories and memories from Ted Henter – Freedom Scientific Blog blog.freedomscientific.com/fsc…
This month, we explore the use of Web verbosity settings to tame excessive speech on some Web site with JAWS. Then, Bryan Vuong introduces us to Innosearch, an innovative approach to shopping. And …
In April, we in the #LibreOffice community attended events in-person (giving away cool merchandise!). We also gave talks online, and announced the location of 2025's conference: blog.documentfoundation.org/bl… #foss #OpenSource
Here’s our summary of updates, events and activities in the LibreOffice project in the last four weeks – click the links to learn more… We started the month by posting a video from Document Freedom Day celebrations with the Nepalese LibreOffice commu…
Featuring - The NVDA 2025.1 Beta - What’s New in 2025.1 - Updated NVDA Expert Certification 2025 - Gene Empowers New Zealand - Changes for Developers - NVDA Add-ons and API Breaking Changes
I just wanted to let you know that this post is not showing up on the nvaccess.org/news page. I like to check that page for an overview of recent happenings at NV Access and I just wanted to let you know about this. Thanks!
And in News: nvaccess.org/news/ (the "latest news" shows the release info, but below that in "recent news", "In-Process 30th April 2025" is the newest item. Unless you are looking somewhere else?
"Microsoft CEO says up to 30% of the company’s code was written by AI."
It can't be 30% by plain math. Just replacing 30% of existing code with new code takes (many) years, and then we include all produced code. If they ONLY used AI to write all code for the last few years, and they wrote it at a high pace, it could *perhaps* be done.
We all know that AI can't write code that good. But sure "up to 30%" could also mean "2%".
@emmadavidson my experience with AI code is similar to that: you can get a lot of suggestions, but you need to filter out the crap and cherry-pick the goodies to be productive. It's very much like having a super-eager junior dev throwing out ideas and code non-stop.
My thought was that it’s all code, and that it includes tons (perhaps a majority) of things like old boilerplate generated from templates and stuff like that.
So… a company that wants you to buy subscriptions to their AI powered tools is over exaggerating the benefits of their AI powered tools? I’m shocked, SHOCKED, to learn that there’s cheating going on…!
Considering that the "AI" cut&paste machine was mostly trained on software copyrighted by other people, including GPL and similar licenses, we are one judge who doesn't believe in the imaginary "AI exemption" away from Microsoft having to cease and desist distributing any such software, with their own CEO confessing to breaking copyright law.
If I understood another company's metric correctly it goes like this: If code has been suggested and a portion of the code made it into the source tree, it counts. But if it's 30% of the code checked in has been "touched" by AI code completion, it'd also be hilariously low for a good product.
I wonder if they count rejected merges from half-working slop bots...
Although my experiences with MSTeams, O365 and SharePoint are so degrading in the last couple of months (not that Teams was any good even before) that I'm open to believe that they actually accept that garbage without much thinking just so they can market it that way
talking to a friend at a different company who works under the Microsoft umbrella, they definitely are pushing hard on the LLM stuff - forcing devs to self-report telemetry on LLM-plugin usage.
Hello, if you care about Servo, the alternative web browser engine that was originally slated to be the next gen Firefox but now is a donation-supported project under the Linux Foundation, you should know of this week's Developments.
Specifically Servo announced a plan to, as of June, start allowing Servo code to be written by "Github Copilot":
Servo’s Technical Steering Committee is looking for feedback regarding our policy for contributions made with generative AI. The current policy, adopted in June 2024, is documented in the Servo boo...
There has been a public comment period on this plan, with public comment so far overwhelmingly negative. I'm not sure how this response will impact the plan; the announcement said enactment was "subject to a review of your feedback", whatever that means.
Personally I think the plan is a horrible idea on both technical and moral grounds, and would end Servo's current frontrunner status as the best candidate for a browser that truly represents the open web and respects the needs of its own users.
Please do not respond to this thread to talk about Ladybird. This is a personal request from me. Just… please. Don't make me have a conversation about Ladybird today. Some other time
This question, both eerie and thought-provoking, provided the core of @bagder 's talk during the @foss_north conference. The Academy Chair's presentation reflected on the progress made with open source, where we are now, and what lies in the future of open source.
it makes me really curious how they measure this. I seriously doubt this number and consider it just AI sales talk. They have too strong incentive to blow up the number.
@maswan I would expect just about every MS dev to have AI enabled as they are probably more or less forced to, and probably are being measured on that metric.
@bagder Ah, maybe I should read the actual quote instead of just the top bullet points too: "I'd say maybe 20%, 30% of the code that is inside of our repos today and some of our projects are probably all written by software," which is both a firmer statement on all code in repository and not just this year's code. But also includes any kind of software generated code.
Assuming of course that it is anything but a wild guess for marketing purposes.
@bagder Some years ago I was at a company where the middle management had promised their boss 100,000 views of their latest product teaser video. The solution was to have it autoplay on most of the site's popular pages. Instant success (as long as you count views as "the video started"). Since then I've had a hard time trusting success numbers from people with a vested interest. A year or so later that whole Facebook pivot to video thing was exposed as an inflated fraud too.
Ondřej Caletka
in reply to Ondřej Caletka • • •May proves to be the cheapest month.
#electricitypricing
Jiří Eischmann
in reply to Ondřej Caletka • • •