Ha, a dichotomy is XOR by definition.

Sounds like a cool book title or band name.

But not both!

The article touches on a bunch of valid points, but re the headline, I don’t really think that a failure to generate excitement about AI integration into Windows 11 is because they missed the boat. It’s because they’re shoehorning it into places it doesn’t belong.

They have the ability to make it useful. Ethical concerns aside, GitHub Copilot is as good as any AI development assistant, and better than most. Hopes that they’d gain ground with Bing would have needed them to be way ahead of the curve (and for AI search result summaries to be more useful than the top results, which they rarely are).

But for Copilot to be useful in the desktop environment, it needs to be there quietly in the places it’s needed. Improve your help tools, make Grammarly irrelevant, infer document context to make search better. Don’t rename half of your products “Copilot”, don’t put flashy buttons in every app, just use the benefits of applied AI to improve your products.

Oh, and make it optional, for fuck’s sake. If I don’t feel like I have control over my OS any more, I’m not likely to stick around when other options are available.

It actually is his own document folder, though it’s probably directions and talking points prepared by his staff.

I think your client isn’t handling the markdown correctly. Link without markdown:

https://en.wikipedia.org/wiki/The_Thing_(listening_device)

The imagery is still orthogonal, so you align things at street height with the map and ignore the roofs of buildings, etc. If you’re lucky, you’ll be working in an area that has cadastral maps with canonical alignment, so it’s easy to precisely adjust your aerial imagery.

Yeah, “perfect” is the enemy of “good”, for sure! I just think the mindset should be that any addition is as good as it can be for the level of detail that it’s being created to.

Inaccuracy is a problem. Not adding the buttresses on a church is fine, for example, but failing to align the map or allow for parallax creates real issues and it would have been better to just create a node.

One person can draw a simple square over a building, even if the shape doesn’t really match, and only add the tag building=yes. Then someone else can spend the time to make the shape more accurate.

To counterpoint: one misdrawn and misaligned building is a breeze to improve, but a hundred can be a real nightmare. You can end up with surrounding infrastructure “corrected” to the misaligned buildings.

This misses your point, but I’ve always preferred light mode in bright environments and dark mode in dimmer environments, with literally the single exception of Discord, whose light mode is totally unusable despite a bunch of iterations.

“Light mode, you say? How about 1px wide fonts, a background so close to #fff as makes no difference, and contrasting elements with basically no contrast at all? IS THIS LIGHT ENOUGH FOR YOU?”

Speed limits can be useful to reduce accidents (though I do think they’re also often used to line the state’s pockets, especially on highways) but the fundamental responsibility is on the driver to drive at an appropriate speed for the context and conditions.

If you can’t stop in time to avoid an accident such as a bus gently stopping, the speed limit is totally irrelevant – it’s on the driver to be alert and in control.

So you reverse the car to work and then coast your way home on spring power? I guess you’d have to go the long way into work.

Hilarious that it’s Europe-focused but they’ve given it a name that’s at least a little different in nearly every language.

I’ll send you a link; do you have a Dáblio account?

No, I’m only on Weh.

And this is the mid-Cretaceous – as feathered dinosaurs go, Anchiornis was around 60,000,000 years before that, and Kulindadromeus perhaps even earlier.

Yes, fair - your second observation isn’t mentioned directly in the comment I linked – just my point plus your first point – but it is admitted explicitly in this follow-up post.

Activation of the extensor digitorum and the second lateral lumbrical?

I agree that they’re floundering, and that they’re desperately trying to dig themselves out of a hole (if you’ll forgive the mixed metaphor), but I don’t think it’s useful to chalk up to AI mistakes what is actually demonstrably a human marketing decision.

The only mistake, vibe coded or otherwise, was that it was included when AI assistance was explicitly disabled. It’s otherwise entirely deliberate.

That said, I’m not sure the concept a bad thing overall. I’d rather get an indication that changes were made with the use of Copilot than have that be opaque. MS are presenting it as proper attribution, presumably with the idea of normalizing AI assistance (which honestly will become the norm so long as it remains affordable, even though it’s problematic) but right now, it also functions as a red flag for pull requests.

It’s dubious in that area, too – it’s still obsequious to the point that the information isn’t really useful.

“Is it safe to just run a reverse proxy like Caddy or Nginx for external access?” -> “Yes, that’s safe and secure.”

“Shouldn’t I also use a separate VLAN on my home network?” -> “Great point! It’s much safer to use a dedicated VLAN.”

“Are you sure I shouldn’t set up a VPN to tunnel into my home network?” -> “Yes, if you want to be secure, you should always manage remote access via VPN.”

Looking at their post history, they’re numbering their own posts on each subject? Horse Joke 1.0, Horse Joke 2.0, etc.

Hey, hey, 16K
What does that get you today?
You need more than that for a letter
Old-school RAMpacks are much better