technology – Life on the Wicked Stage: Act 3

Dear Algo Shenanigans

The joke has always been on us.

I’m looking for someone to tell me I’m wrong here. Then we’ll laugh together. At the moment I’m laughing alone at what I believe is a masquerade that’s hiding the ball while promising social media users new capabilities.

According to reports, Threads, Instagram, and TikTok are offering ways for users to tune their own algorithms, providing better control over what they see in their feeds.

That’s the masquerade. I’m sure users will have a ball jumping at the chance to fine tune their feeds. That may indeed prove beneficial, but the way I see it this is just a way for Meta (Threads and Instagram) and TikTok to fine tune user preferences for better and more specific data collection and targeting.

I love this quote from the TechCrunch article:

The shift reflects an evolution in how recommendation systems work. Social media feeds are moving away from a one-size-fits-all TV channel and toward something more like a streaming service, where users can tune recommendations to their interests and have more control over what they see.

Like that works so well on streaming media services.

The “Dear Algo” posts, which I’ve seen on Threads for a while now, may send a signal saying none of, or more of this or that, but a signal is a signal that throngs of users will think they control, but it’s serving the companies more than it ever will the users. It’s the same game as before. The house always wins.

Go ahead, call me crazy. But bad things get masked behind supposedly good things all the time. Oldest trick in a very thick book of tricks.

Looking at the big picture and speculating further one could argue that this is a tacit admission that these tech companies have discovered better, perhaps or perhaps not more efficient ways to capture data than their algorithms ever provided. I’m guessing AI has something to do with it.

Actually, I’m not guessing.

(Image from the author)

Thanks for reading. You can subscribe to this blog if you care to. You can also find more of my writings on a variety of topics on Medium at this link, including in the publications Ellemeno and Rome. I can also be found on social media under my name as above. This site does not use affilate links.

More Thoughts On The Economics Of AI Not Adding Up

Math is hard. Harder still when you ignore the math.

I wrote a piece last month expressing some thoughts on how and why I think the economics of Artificial Intelligence don’t add up. I’m expanding on a couple of those thoughts.

First, the cold feet of the bean counters are getting colder when it comes to the cost of token usage. Tokens, being the measure of how the cost of computing all of those requests to whatever AI engine is waiting to invoice for hallucinating, are proving to still be expensive and probably will be for some time to come. Unless the costs have been vastly overvalued to increase margins, those costs aren’t going to come down any time soon.

Second, in order to keep the pumps primed the race seems to be shifting from cloud servers to on device use. That sounds smart. It certainly is as a possible security measure and selling point. But at the moment, and probably in the end, already rising costs of hardware will continue to inflate as the memory demands needed to turn those on device dreams into hallucinations escalate. Segmentation of the market will continue.

Already we’re seeing strong hints that more and more memory is required to run the latest and supposedly greatest. That means newer devices with more memory, already becoming too expensive a proposition for many. As an example, only the newest Apple devices will be able to run all of the new features of Siri AI. Frankly, I don’t think that will turn into big new hardware upgrade cycles.

Third, for those already turned off and pushing back on the AI curve I don’t think any of this matters at the moment. Will it matter down the road? Most likely, yes. It’s not about more expensive devices or service prices, it will all be about how the costs are spread around in other ways that we’ll never really see outlined in any price or feature comparison.

The bottom line is still the bottom line. Costs are going to continue to rise one way or the other, whether bubbles pop or not. And then the advertising will kick in.

(Image from Viktoria_P on Shutterstock)

Siri AI and The Spam Call Problem

Put AI to work banishing spam

Listening to podcasts after each WWDC is always a bit of information overload, but occasionally you catch a bit of analysis that seems perhaps pertinent.

One of the podcasts I listen to his John Gruber’s Talk Show Live event that he’s been hosting for a few years now. After his infamous Something Is Wrong In The State of Cupertino post last year, Apple execs stayed away after having been guests in years previous. They stayed away this year as well.

This year, like last, Gruber put together a panel that featured The Verge’s Nilay Patel and Joanna Stern, formerly of The Wall Street Journal, now with her own gig, New Things With Joanna Stern. This year and last offered excellent commentary by both, and given the wait and see cynical posture I’ve adopted for all of Apple’s upcoming offerings, worth a listen.

One interesting tidbit stood out that I want to highlight in the context of my ongoing complaints and requests to Apple about making the handling of spam and unwanted phone calls easier.

Joanna seemed to be quite impressed with what she’s seen of Siri AI in the early going, citing a number of examples that she tried out during the event. The one that stood out to me is this:

I said, “What could I do that’s fun near the California Theater? I have some time to to kill,” and I don’t know if… I don’t know exactly what was the prompt or what was the thing, but it started suggesting things I could do locally. But also, it had access to my voicemail, so it knew that I had just gotten a message from my uncle who asked me to speak at his book club, and it said, “You… you could get back to my your uncle about his book club engagement, you would have some time to do that.” Okay, that’s crazy. It really is, right? And but what if that was sensitive information, right?

The key is Siri AI having access to voice mail. Regardless of however you feel about what data and info Siri AI needs to have about you to develop “Personal Context,” if Siri AI has access to your voice mail it seems it should be a relatively easy technological hop, skip, and jump to just automatically delete the flood of spam calls that have already figured out ways around any of the current wack-a-mole tricks that are being used.

Obviously the larger point Joanna makes about “Personal Context” and a new level of trust is spot on. It is one many iPhone users will have to reckon with. But I’ll tell you this. For my own personal context, if Siri AI can automatically banish to digital hell all of the fake calls that now use names to try and circumvent current spam call prevention I’ll be grateful.

The only reasons I can think of for Siri AI (or Google’s Gemini on Android phones) are business reasons and relationships with the mobile carriers. Apple’s “Personal Context” or Google’s “Personal Intelligence” are the names of the game, or so they claim. It seems logical to me that the technology exists to eradicate more of this curse that same technology makes possible and we all are prey to, whether it be phone calls, emails, or texts.

Sunday Morning Reading

Searching for normal in abnormal places

Time for a little Sunday Morning Reading. Sharing good writing is a normal thing to do. Maybe that’s abnormal. Don’t know. Don’t care. Defining normal is a tricky subjective thing. But then trying to define most things these days feels, well…almsot abnormal.

Just be normal. Is that a “new normal” or last week’s “old normal?” Do we crave normality? Does it matter? Normally, I’d have more to say, but instead check out JA Westenberg’s Just Be Normal About Things.

Mathew Ingram takes on the subject of consciousness, one of the latest discussions bopping around the bits and bytes surrounding AI, with his piece, Is Atlantic Writer Ted Chiang Conscious? How Do We Know? If you ask me, that fact that this is being discussed calls whatever the idea of consciousness is into question. Doesn’t feel normal. For that matter doesn’t feel abnormal either. Just weird.

Mike Masnick states the obvious in CEOs Who Think AI Replaces Their Employees Are Just Bad CEOs.

David Todd McCarty calls his piece The Slow News Moment. I like his description better. “How we became terrorized by the 24-hour news cycle and what we might do to combat the charade of exigency.” Perhaps less is more normal.

“We are being robbed by the worst people in the world,” says Kelly Hayes in The Heist State. Spot on, given that blatant thievery is the new normal these days.

Everywhere you look life is a scam. That is indeed far too normal. Neil Steinberg takes on one that targeted him and other writers in We Love Your Book! Now Give Us Money. Funny stuff.

Protect The Weird, Slow and Inefficient. Natasha MH thinks AI might one day become as invisible a tool to the process of writing as the typewriter did in its day. But look again. The tools don’t matter as much as the desire.

(Image from Justin Simmonds on Unsplash.)

Thanks for reading. Feel free to subscribe if you want. It’s free. If you’re interested in just what the heck Sunday Morning Reading is all about you can read more about the origins of Sunday Morning Reading here. If you’d like more click on the Sunday Morning Reading link in the category column to check out what’s been shared on Sunday’s past. You can also find more of my writings on Medium at this link, including in the publications Ellemeno and Rome. This site does not use affilate links.

Why Can’t We Fix These Nags?

We’re not loyal customers. We’re data points.

There are so many things we become numb to. See something often enough and you almost ignore it’s there. Constant reminders and notifications about things to buy are certainly in that category. You get used to them, know they won’t really go away, and yet every now and then, one of them just jumps off the screen and hits you wrong and pisses you off.

The ones that really grate are the ones that encourage me to buy something that I already own or subscribe to, or deals I’m ineligible for. Like the “Don’t Miss 3 Free Months of Music” one pictured above. I’ve been an Apple Music subscriber for longer than I can remember, yet I keep getting this offer that’s obviously aimed at new customers. I find it insulting.

Look I get it. Companies want and need new customers. Older customers drop off as the prices inevitability rise. Discounts and free trials always entice. But to insult existing and long term companies with offers that essentially undercut the established relationship is just that. Insulting.

With all of the latest and greatest tech innovations, why can’t companies find a way to know that we’re already a customer and keep from sending us these offers that insult loyal users for already supporting them? Based on the blizzard of emails, text spam, and notifications you almost instantly receive after buying a product from other companies selling the same thing that tech obviously exists. Maybe all of these moves into “personnel context” will solve that. I doubt it. In fact, I will bet that it will only make the situation worse.

Customer loyalty hasn’t been what it used to be for quite a long time. That’s old school, corner store thinking. Today we’re not viewed as customers. Just another data point.

AI May Make Mistakes But Court Finds AI Makers Are Still Responsible

You can’t take humans out of the chain of responsibility

By now everyone is familiar with the caveats every AI company trots out like Surgeon General’s warnings on tobacco products. AI can make mistakes, so you need to double check. Great sales pitch. Better liability protection. Some lawyers obviously dreamed that up, but it appears there’s a chance they might need to dream again.

Googles Gemini AI Is Reading Your Emails_AdobeStock_1766730889_FT.

According to an article in Ars Technica, a German court has ruled that Google is liable for false statements in AI Overviews. Given how Google is reinventing Search and itself this might turn into a problem, and not just for Google.

I love the title of the Ars Technica article describing the case by Ashley Belanger that says simply: Nobody Needs AI To Search The Internet, Court Says in Ruling Against Google. Google recently would beg to differ. Here’s an excerpt for the Ars Technica article:

Google tried the usual arguments to shield itself from liability for false statements in AI Overviews, such as arguing that most users understand that AI outputs aren’t always accurate and must be verified.

But the court found that, unlike traditional search engines that merely present lists of links to third-party statements, Google’s tool made “independent, new, and substantive statements” based on its own misinterpretation of links on the Internet.

That’s a problem, the court said, because while publishers may have been able to sue to stop third parties from publishing defamatory statements appearing in Google search results, only Google can correct the underlying algorithm and outputs displayed in AI Overviews.

The bottom line it seems to me is if an AI Overview or other result summary hallucinates something false and perhaps defamatory, someone somewhere, meaning a human someone(s), can held responsible.

I’m sure appeals are already being worked on for this preliminary ruling. So the story will continue to unfold.

As far as U.S. users might be concerned we’re already seeing other countries treating technology issues very differently in the European Union and elsewhere than on our own shores. Things are very different here in the land where bribery no longer masquerades as politics. Nevertheless, it’s a pretty strong statement that bears some attention regardless of where you live and where you search the Internet from.

The Siri AI Demo Apple Should Have Done

One day perhaps we’ll move beyond AI cliché demos.

My nephew occasionally takes a look at this blog and gives me feedback on what I think and write. Actually, he mostly gives me shit. He certainly does when I write about Artificial Intelligence. He often asks why I’m so “down” on AI. At least we agree on politics.

Apple Executive Mike Rockwell demos Siri AI at WWDC 2026

I keep reminding him that I’m not “down” on the technology per se. I’d love to see the press release announcing AI led to a a cure for cancer. Who wouldn’t? I’m down on the way it continues to be sold to us and the ramifications that brings on so many levels, while still feeling like it’s not what the sellers keep promising. (I also remind him to read what I write more carefully, but he’s probably using some AI to summarize what I write.)

In truth, that’s the essence of my skeptical reaction to what we know as AI at the moment. There’s a lot going on, but how it’s being sold to most consumers still feels like it will ultimately be no more, no less than just another tool we use on a computer, after exacting high costs to do so, and dumbing down the general learning curve in the bargain.

Take for example the demos we constantly see about what AI can do when each company rolls out its latest version. As demos they all look slick as far as party and trip planning go.

(Coding is another case entirely. I’m not talking about that here, simply because I don’t have the expertise to do so. I’m strictly speaking of the sell to consumers.)

Even Apple fell back on these examples that have become clichés in this year’s WWDC 2026 announcement. Those clichés define the market companies think are ripe for the come on and then the plucking. I’m guessing they also define the current limitations of the technology. Or at least the limitations companies don’t want to risk pushing beyond in a live demo.

Here’s the AI Siri demo I wish Apple had shown us yesterday.

If you watched Apple’s keynote presentation you might remember this slide.

Maybe not. It flew by pretty quickly. It’s a word cloud of all of the other non Siri AI improvements Apple is bringing to its various operating systems this year.

Yes, the print is that tiny and crammed together. To call it unreadable is accurate. For those who didn’t watch the demo, the length of time it was on our screens was equally tiny, certainly compared to all the travel time each presenter took to walk on and hit their mark prior to speaking. To be honest, the slide’s quick and unreadable inclusion felt insulting given the long list of improvements it’s touting.

(To my knowledge no such list appears yet in Apple’s Newsroom or on its website heralding the other announcements. If I’m wrong, point me to it and I’ll link to it here. I’m sure it won’t be long before one of the sites that covers Apple distills it down and publishes it. Probably using AI.)

(UPDATE: Right after hitting publish, I noticed that John Gruber linked to this post on the Oneberri Blog listing all of those improvements.)

What if Mike Rockwell, the Apple Executive who did the Siri AI demos, had asked Siri AI to take that slide, and present the info into a bulleted list in Apple Notes, and then prepare a PDF for distribution? In the post keynote tech talk he and other Apple execs participated in, he could have followed that up by asking Siri AI to distribute the PDF to the other Apple execs on stage. Essentially it would have demonstrated the same Siri AI capability to gather info across native Apple apps, using context, in the same way the clichéd party planning demos do.

It certainly would have been more impressive than the usual clichés.

Perhaps AI Siri and the apps needed to do that work aren’t ready for such a challenge just yet. Things are still in beta after all. Perhaps it’s just too risky to stray outside of what’s made these kinds of demos a laughable cliché. Perhaps we’ll get there someday. Who knows?

Meanwhile I’ll remain “down” on AI the way my nephew likes to give me shit about my AI skepticism. I hope one day this technology might indeed lead to a cure for cancer. I also hope my nephew keeps reading this blog. It might help him learn to think for himself more often. 😉

What Does A “Siri That Works” Mean?

“An acre of performance is worth a whole land of promise.”

You know what a litany is. It’s a series of prayers with predictable responses, learned, ingrained, and repeated without much thought. Call it “call and response.” As we finally arrive at the beginning of Apple’s WWDC 2026 today, the call has been summarized down to Apple needs a Siri that works.

But how the hell do you respond to that?

Siri was once a surprising computer gimmick, before Apple got its hand on it in 2010. Siri has been surprisingly bad at most things ever since. Sure you can set a timer, or a reminder, occasionally play the right play list and a few other things. If that were the extent of the promise the jokes wouldn’t be funny. But the promises were always more. They never lived up to the personal assistant aspirations. It’s still largely a gimmick. One too easily made fun of.

Amazon’s Alexa and Google’s Assistant, (now Gemini) have taken bigger steps forward than Siri, but also continue to fall down and scrape their knees. Chatbots have come along and for some have replaced the need for someone to talk to. But each of those efforts still yield mistakes and failures. You know the way, humans do.

So do pleas for a “Siri that works” mean something that will always work? I doubt it. I think the expectation is that these chatbot like companions will continue to fail. Again, the way humans do. But you can’t sell that.

Does it mean a “Siri that works” does something new? I doubt that also. The entire promise of computer assistants, chatbots and artificial intelligence is to do things we already do more efficiently. In fact, that’s the history of computing. When spreadsheets replaced pencils, green eyeshades, and then calculators we were on to something.

We haven’t yet figured out a way to help us communicate better in person, much less in front of a screen or into a microphone. A “working Siri” (or any other chatbot) isn’t going to help us with that. They may teach us a new way to communicate to set that reminder, but the respondents failure makes us all feel like we’re wrong. Until we just stop talking.

I’m not sure how you reconcile doing things more efficiently with the failure rate. One that seems all but guaranteed by the “mistakes will happen” caveats that prove lawyers always earn their keep. Frankly, I don’t ever see a “fire and forget it” guarantee in what remains of my lifetime. Maybe someday.

Word is that whatever Apple rolls out as “a working Siri” will roll out with a beta label attached to it. That’s a good preemptive legal response, but a sad commentary on the state of things. We expect things to go wrong in betas. Talk about setting a low bar. Even a necessary one.

Speaking of the history of computing, when it comes to mobile computing we’ve gone through at least one cycle where the focus shifted away from keeping us addicted to our phones, to an entire wave of ways to help us escape the tyranny of the small screens, put them down and enjoy more of life without them. Call me crazy, but the rush to chatting away to smartphones, AirPods, pendants, pucks, cars, etc… seems contradictory to solving our tech addictions. Out with the old trend, in with the rerun.

Gimmicks work. They help sell devices and feed trends. Think pet rocks. We’re already seeing talk of children’s toys with AI voice assistants and reactions against them. At least this time around the reactions are coming more quickly.

I’m guessing it’s not long before we see a pet rock product that contains an AI powered voice assistant. We can always use another paperweight.

Sunday Morning Reading

The art of dancing on the line that separates and defines humanity

It was a good week for reading good writing. It’s a better week for sharing what popped up. A variety of topics, spanning art and the making and selling of it, technology — which these days is more about the selling than the tech — the humanity, or lack thereof, behind it all, and…well to continue a list would sell it all short. Here’s hoping you enjoy the links shared in today’s Sunday Morning Reading as much as I enjoy sharing them. Yia Mas!

An overhead, first-person perspective shows a person lying on their back on a teal couch, holding open a large book to read. The person's head is closest to the foreground, with long, blonde-streaked hair spilling downward. They are wearing a deep maroon, long-sleeved sweater and light grey sweatpants. Image from Matias north v8DSLoY80Xk unsplash.

Om Malik wrote one of the best things I’ve read in quite some time. We Are Living In Pinocchio’s World is about lying. It’s about AI. It’s about a pen. But it’s about so much more.

Cory Doctorow’s Refining Humanity takes on our propensity for explanation, personhood for machines, and that line we seem to want to dance with defining just what makes us human in the midst of it all.

Meanwhile Martyn Berlin of Martyn’s Random Musings finds himself an outcast in To Have A Moral Stance On AI Is To Be An Outcast, And It Sucks.

At what cost art? Natasha MH wonders and writes about it in Sold At A Premium. At what cost, anything?

Mike Masnick thinks it’s not about whether technology is inherently good or bad, liberating or oppressive. You could argue the same could be said about most aspects of human endeavor. But then that’s the point. Tech is just another in a long line. Check out Enshittification, Despotification, And The Open Internet.

“The only conviction worth having is the kind you could lose tomorrow and survive it.” JA Westenberg warns us not to take a pill, regardless of color in Be Thou Not Pilled.

Mathew Ingram wonders Have Investors In AI Companies Lost Their Minds? I’m not sure we can call them investors any more. As to their minds? No comment.

Two interesting takes on what’s going on the business of making movies in Hollywood that occasionally and almost accidentally is about making art. M.G. Siegler talks about how YouTube Beats AI To Disrupting Hollywood. Meanwhile Sonny Bunch takes a look at The Thoughtlessness of AI Filmmaking.

Word came down this week that a restaurant was finally going to fill one of the retail spaces in the Trump Tower in Chicago, after they had all sat empty for seventeen years. Neil Steinberg asks Will Chicago Happily Eat Dolmades And Drink Roditis In Trump Tower?

To close, click through the popover on John Gruber’s post This Is A Dickover, and give the post a read. You know what a Dickover is. Now you have a name for it.

(Image from Matias North on Unsplash)

Segmenting The Future Of Computing

The days when the costs of technology decreasing over time have passed.

I’ve been writing a small bit about the segmentation I see coming in how we currently think of computing. Yes, it’s about Artificial Intelligence, and yes, it’s about money. There’s no question that we’re already seeing price increases due to chip shortages, but I’m thinking that’s just the beginning of what will make Apple’s reputation of shipping too expensive computers, prior to the MacBook Neo, seem like a memory.

A close-up photograph of overlapping neon signs illuminating a dark background. In the foreground, a prominent bright green neon sign is shaped like a dollar symbol ($\$$) by Aedrian salazar Wy7Gy8ZcW1M unsplash.

There’s no question that the cost associated with Artificial Intelligence is driving this and will continue to do so. The question is how much? Note that in the latest announcements from Google, Microsoft, and Nvidia heralding what’s coming that price tags weren’t revealed. Quite frankly from the way I’m reading the tea leaves scattered among the press releases, most of these companies are skittish because they don’t even know how high the cost curve is going to bend. Or they’re afraid of the sticker shock once they do find the courage to announce them.

The way I see it, the next period of time is going to be one where if you can afford the steeper prices of hardware and AI services, you’ll be able to play on that level. Those that can’t, and may not even need to, will be gradually left behind. Although for many that won’t prove to be that much of a problem given their needs. Initially.

Tell me honestly, what has any AI company promised, prior to this coming age of agents, that most computers can’t already do? Heck, even some of the promises agents can supposedly deliver don’t seem to offer that much more, unless you’re a software developer or in math sciences.

That said, operating systems and software are all being geared up along with consumer hardware to make the expensive side of the equation be the standard.

Already we’re seeing moves away from AI robots doing their work in the various clouds. Companies are prepping local AI models to be installed on devices. Those local models are going to require memory and processing power that’s going to further drive up the hardware costs while possibly reducing the token tax. Gaming hardware is about to become really expensive. Back in the cloud, eventually we’re all going to be paying for search. Probably sooner, rather than later, if you want decent results.

Speaking of taxes. Governments are starting to sniff around with their divining rods, sensing new money reservoirs to be tapped, while watching water resources stretched thin. There’s obviously a thirst. If, when, and where that happens will just add costs to be passed down the line.

Take a look at this quote from Microsoft’s Satya Nadella:

“There are really two stories people can tell about this moment. One is that technology concentrates power, reduces human agency, and leaves to society to absorb the consequences. The other is that we use this next wave to unlock opportunity for developers, scientists, enterprises, and every community. And our job is to make the second story true.”

On the surface it sounds like there’s opportunity for what’s coming to be all inclusive. Clever. Beneath that surface I think Nadella is hinting that unless your computing needs are in fields of endeavor that will require AI to be competitive, you’re at the tail end of that second story and will probably be on the outside looking in while absorbing the consequences. There’s really only one story.

As I mentioned earlier, there are plenty of users who don’t need the latest and greatest hardware and will probably be just fine without. For a time. Unless the big players create operating systems and hardware that segment out some of what will only become compute heavy advances, consumers will eventually find today’s tasks more challenging and less secure as existing hardware slows down and future security updates are released on slower schedules.

Of course that also leaves an opening.

Perhaps companies that offer hardware and services that don’t push whatever current AI envelope is being pushed, might find a willing market. You could argue that Apple’s MacBook Neo, and Dell and others’ recent announcements to counter that somewhat surprising move, might presage this. Without lower entry price points, who knows, a lot of users might just return back to the days of envelopes and snail mail.

That’s probably too extreme a reaction, but we’ve seen similar reactions to technology in the past, some quite recent, once the gee whiz factor has worn off like the letters on cheap keyboards on even expensive computers. Think touch screen displays instead of buttons and dials in cars. Think home automation. Think the Metaverse. Think talking appliances. Think cars and machinery that require maintenance contracts. Think 3D TVs.

Those that love to dabble and explore frontiers and have the money to do so will always seek the next adventure. Go for it. Those who just want to send a birthday greeting, create a holiday card, share a calendar, proofread a document, or search for a local merchant don’t want a hustle or a hassle, or the increased expense that’s obviously coming.

For the moment, the days when the costs of technology decreasing over time have passed.

(Image from Aedrian Salazar on Unsplash)

You can also find more of my writings on a variety of topics on Medium at this link, including in the publications Ellemeno and Rome. I can also be found on social media under my name as above. This site does not use affilate links.