AI Produces Human Clipboards
July 16, 2025
![Dino 5 18 25_thumb[3]](https://www.arnoldit.com/wordpress/wp-content/uploads/2025/07/Dino-5-18-25_thumb3-1.gif "Dino 5 18 25_thumb[3]")
No smart software involved with this blog post. (An anomaly I know.)
The upside and downside of AI seep from my newsfeed each day. More headlines want me to pay to view a story from Benzinga. Benzinga, a news release outfit. I installed Smartnews on one of my worthless mobile devices. Out of three stories, one was incoherent. No thanks, AI.
I spotted a write up in the Code by Tom blog titled “The Hidden Cost of AI Reliance.” It contained a quote to note; to wit:
“I’ve become a human clipboard.”
The write up includes useful references about the impact of smart software on some humans’ thinking skills. I urge you to read the original post.
I want to highlight three facets of the “AI problem” that Code by Tom sparked for me.
First, the idea that the smart software is just “there” and it is usually correct strikes me as a significant drawback for students. I think the impact in grade school and high school will be significant. No amount of Microsoft and OpenAI money to train educators about AI will ameliorate unthinking dependence of devices which just provide answers. The act of finding answers and verifying them are essential for many types of knowledge work. I am not convinced that today’s smart software which took decades to become the next big thing can do much more than output what has been fed into the neural predictive matrix mathy systems.
Second, the idea that teachers can somehow integrate smart software into reading, writing, and arithmetic is interesting. What happens if students do not use the smart software the way Microsoft or OpenAI’s educational effort advises. What then? Once certain cultural systems and norms are eroded, one cannot purchase a replacement at the Dollar Store. I think with the current AI systems, the United States speeds more quickly to a digital dark age. It took a long time to toward something resembling a non dark age.
Finally, I am not sure if over reliance is the correct way to express my view of AI. If one drives to work a certain way each day, the highway furniture just disappears. Change a billboard or the color of a big sign, and people notice. The more ubiquitous smart software becomes, the less aware many people will be that it has altered thought processes, abilities related to determine fact from fiction, and the ability to come up with a new idea. People, like the goldfish in a bowl of water, won’t know anything except the water and the blurred images outside the aquarium’s sides.
Tom, the coder, seems to be concerned. I do most tasks the old-fashioned way. I pay attention to smart software, but my experiences are limited. What I find is that it is more difficult now to find high quality information than at any other time in my professional career. I did a project years ago for the University of Michigan. The work concerned technical changes to move books off-campus and use the library space to create a coffee shop type atmosphere. I wrote a report, and I know that books and traditional research tools were not where the action was. My local Barnes & Noble bookstore sells toys and manga cartoons. The local library promotes downloading videos.
Smart software is a contributor to a general loss of interest in learning the hard way. I think smart software is a consequence of eroding intellectual capability, not a cause. Schools were turning out graduates who could not read or do math. What’s the fix? Create software to allow standards to be pushed aside. The idea is that if a student is smart, that student does not have to go to college. One young person told me that she was going to study something practical like plumbing.
Let me flip the argument.
Smart software is a factor, but I think the US educational system and the devaluation of certain ideas like learning to read, write, and “do” math manifest what people in the US want. Ease, convenience, time to doom scroll. We have, therefore, smart software. Every child will be, by definition, smart.
Will these future innovators and leaders know how to think about information in a critical way? The answer for the vast majority of US educated students, the answer will be, “Not really.”
Stephen E Arnold, July 16, 2025
An AI Wrapper May Resolve Some Problems with Smart Software
July 15, 2025
No smart software involved with this blog post. (An anomaly I know.)
For those with big bucks sunk in smart software chasing their tail around large language models, I learned about a clever adjustment — an adjustment that could pour some water on those burning black holes of cash.
A 36 page “paper” appeared on ArXiv on July 4, 2025 (Happy Birthday, America!). The original paper was “revised” and posted on July 8, 2025. You can read the July 8, 2025, version of “MemOS: A Memory OS for AI System” and monitor ArXiv for subsequent updates.
I recommend that AI enthusiasts download the paper and read it. Today content has a tendency to disappear or end up behind paywalls of one kind or another.
The authors of the paper come from outfits in China working on a wide range of smart software. These institutions explore smart waste water as well as autonomous kinetic command-and-control systems. Two organizations funding the “authors” of the research and the ArXiv write up are a start up called MemTensor (Shanghai) Technology Co. Ltd. The idea is to take good old Google tensor learnings and make them less stupid. The other outfit is the Research Institute of China Telecom. This entity is where interesting things like quantum communication and novel applications of ultra high frequencies are explored.
The MemOS is, based on my reading of the paper, is that MemOS adds a “layer” of knowledge functionality to large language models. The approach remembers the users’ or another system’s “knowledge process.” The idea is that instead of every prompt being a brand new sheet of paper, the LLM has a functional history or “digital notebook.” The entries in this notebook can be used to provide dynamic context for a user’s or another system’s query, prompt, or request. One application is “smart wireless” applications; another, context-aware kinetic devices.
I am not sure about some of the assertions in the write up; for example, performance gains, the benchmark results, and similar data points.
However, I think that the idea of a higher level of abstraction combined with enhanced memory of what the user or the system requests is interesting. The approach is similar to having an “old” AS/400 or whatever IBM calls these machines and interacting with them via a separate computing system is a good one. Request an output from the AS/400. Get the data from an I/O device the AS/400 supports. Interact with those data in the separate but “loosely coupled” computer. Then reverse the process and let the AS/400 do its thing with the input data on its own quite tricky workflow. Inefficient? You bet. Does it prevent the AS/400 from trashing its memory? Most of the time, it sure does.
The authors include a pastel graphic to make clear that the separation from the LLM is what I assume will be positioned as an original, unique, never-before-considered innovation:
Now does it work? In a laboratory, absolutely. At the Syracuse Parallel Processing Center, my colleagues presented a demonstration to Hillary Clinton. The search, text, video thing behaved like a trained tiger before that tiger attacked Roy in the Siegfried & Roy animal act in October 2003.
Are the data reproducible? Good question. It is, however, a time when fake data and synthetic government officials are posting videos and making telephone calls. Time will reveal the efficacy of the ‘breakthrough.”
Several observations:
- The purpose of the write up is a component of the China smart, US dumb marketing campaign
- The number of institutions involved, the presence of a Chinese start up, and the very big time Research Institute of China Telecom send the message that this AI expertise is diffused across numerous institutions
- The timing of the release of the paper is delicious: Happy Birthday, Uncle Sam.
Net net: Perhaps Meta should be hiring AI wizards from the Middle Kingdom?
Stephen E Arnold, July 15, 2025
Google Is Great. Its AI Is the Leader, Just As Philco Was
July 15, 2025
No smart software involved with this blog post. (An anomaly I know.)
The Google and its Code Red Yellow or whatever has to pull a revenue rabbit out of its ageing Stetson. (It is a big Stetson too.) Microsoft found a way to put Googzilla on its back paw in January 2023. Mr. Nadella announced a deal with OpenAI and ignited the Softies to put Copilot in everything, including the ASCII editor Notepad.
Google demonstrated a knee jerk reaction. Put Prabhakar in Paris to do a stand up about Google AI. Then Google reorganized its smart software activities… sort of. The wizards at Google has pushed out like a toothpaste tube crushed by a Stanford University computer science professor’s flip flops. Suffice it to say there are many Google AI products and services. I gave up trying to keep track of them months ago.
What’s happened? Old-school, Google searches are work now. Some sites have said that Google referral traffic is down a third or more.
What’s up?
“Google Faces Threat That Could Destroy Its Business” offers what I would characterize as a Wall Street MBA view of the present day Google. The write up says:
As the AI boom continues to transform the landscape of the tech world, a new type of user behavior has begun to gain popularity on the web. It’s called zero-click search, and it means a person searches for something and gets the answer they want without clicking a single link. There are several reasons for this, including the AI Overview section that Google has added to the top of many search result pages. This isn’t a bad thing, but what’s interesting is why Google is leaning into AI Overview in the first place: millions of people are opening ChatGPT instead of Google to search for the things they want to know.
The cited passage suggests that Google is embracing one-click search, essentially marginalizing the old-school list of links. Google has made this decision because of or in response to OpenAI. Lurking between the lines of the paragraph is the question, “What the heck is Google doing?”
On July 9, Reuters exclusively reported that OpenAI would soon launch its own web browser to challenge Google Chrome’s dominance.
This follows on OpenAI’s stating that it would like to buy the Chrome browser if the US government forces Google to sell is ubiquitous data collection interface with users. Start ups are building browsers. Perplexity is building browsers. The difference is that OpenAI and Perplexity will use AI as plumbing, not an add on. Chrome is built as a Web 1 and Web 2 service. OpenAI and Perplexity are likely to just go for Web 3 functionality.
What’s that look like? I am not sure, but it will not come from some code originally cooked up someplace like Denmark and refurbished many times to the ubiquitous product we have today.
My view is that Google is somewhat disorganized when it comes to smart software. As the company tries to revolutionize medicine, create smart maps, and build expensive self driving taxis — people are gravitating to ChatGPT which is now a brand like Kleenex or Xerox. Perplexity is a fan favorite at the moment as well. To add some spice to the search recipe, Anthropic and outfits like China Telecom are busy innovating.
What about Google? We are about to learn how a former blue chip consultant will give Google more smarts. Will that intelligence keep the money flowing and growing? Why be a Debbie Downer. Google is the greatest thing since sliced bread. Those legal actions are conspiracies fueled by jealous competitors. Those staff cutback? Just efficiencies. Those somewhat confusing AI products and services? Hey, you are just not sufficiently Googley to see the brilliance of Googzilla’s strategy.
Okay, I agree. Google is wonderful and the Wall Street MBA type analysis is wonky, probably written with help from Grok or Mistral. Google is and will be wonderful. You can search for examples too. Give Perplexity a try.
Stephen E Arnold, July 15, 2025
Killing Consulting: Knowledge Draculas Live Forever
July 14, 2025
No smart software involved with this blog post. (An anomaly I know.)
I read an opinion piece published in the Substack system. The article’s title is “The Consulting Crash Is Coming.” This title is in big letters. The write up delivers big news to people who probably did not work at large consulting companies; specifically, the blue chip outfits like McKinsey, Bain, BCG, Booz, Allen, and a handful of others.
The point of the write up is that large language models will put a stake in the consulting Draculas.
I want to offer several observations as a former full-time professional at one of the blue chip outfits and a contractor for a couple of other pay-for-knowledge services firms.
First, assume that Mr. Nocera is correct. Whatever consulting companies remain in business will have professionals with specific work processes developed by the blue chip consulting firms. Boards, directors, investors, non-governmental organizations, individual rich people, and institutions like government agencies and major academic institutions want access to the people and knowledge value of the blue chip consulting firms. A consulting company may become smaller, but the entity will adapt. Leaders of organizations in the sectors I identified will hire these firms. An AT Kearney-type of firm may be disappeared, but for the top tier, resiliency is part of the blue chip DNA.
Second, anyone familiar with Stuart Kauffman (Santa Fe Institute) is familiar with his notion of spontaneous order, adjacency, and innovations creating more innovations. As a result of this de facto inferno of novelty and change, smart people will want to hire other smart people in the hopes of learning something useful. One can ask a large language model or its new and improved versions. However, blue chip consulting firms and the people they attract usually come up with orthogonal ideas and questions. The knowledge exercise builds the client’s mental strength. That is what brings clients to the blue chip firm’s door.
Third, blue chip consulting firms can be useful proxies for [a] reorganizing a unit and removing a problematic officer, [b] figuring out what company to buy, how to chop it up, and sell of the parts for a profit, [c] thinking about clever ways to deploy new technology because the blue chip professionals have first hand expertise from many different companies and their work processes. Where did synthetic bacon bits originate? Answer: A blue chip consulting company. A food company just paid the firm to assemble a team to come up with a new product. Bingo. Big seller.
Fourth, hiring a blue chip consulting firm conveys prestige to some clients. Many senior executives suffer from imposter syndrome. Many are not sure what happened to generate so much cash and market impact. The blue chip firm delivers “colleagues” who function to reduce the senior executive’s anxiety. Those senior executives will pay. My boss challenged Jack Welch in a double or nothing bet worth millions in consulting fees regarding a specific report. Mr. Welch loved the challenge from a mere consulting firm. The bet was to deliver specific high value information. We did. Mr. Welch paid the bill for the report he didn’t like and the one that doubled the original fee. He said, “We will hire you guys again. You think the way I do.”
Net net: Bring on the LLMs, the AI, the smart back office workflows. The blue chip consulting firms may downsize; they may recalibrate; they will not go away. Like Draculas, they keep getting invited back, to suck fees, and probably live forever.
Stephen E Arnold, July 14, 2025
Deezer: Not Impressed with AI Tunes
July 14, 2025
Apparently, musical AI models have been flooding streaming services with their tracks. But ai music is the tune for the 21st century is it not? The main problem: these bots can divert payments that should have gone to human artists. Now, we learn from MSN, “Streaming Platform Deezer Starts Flagging AI-Generated Music.” The article, originally published at The Economic Times, states:
“Deezer said in January that it was receiving uploads of 10,000 AI tracks a day, doubling to over 20,000 in an April statement — or around 18% of all music added to the platform. The company ‘wants to make sure that royalties supposed to go to artists aren’t being taken away’ by tracks generated from a brief text prompt typed into a music generator like Suno or Udio, Lanternier said. AI tracks are not being removed from Deezer’s library, but instead are demonetised to avoid unfairly reducing human musicians’ royalties. Albums containing tracks suspected of being created in this way are now flagged with a notice reading ‘content generated by AI’, a move Deezer says is a global first for a streaming service.”
Probably a good thing. Will other, larger streaming services follow suit? Spotify, for one, is not yet ready to make that pledge. The streaming giant seems squeamish to wade into legal issues around the difference between AI- and human-created works. It also points to the lack of a “clear definition” for entirely AI-generated audio.
How does Deezer separate the human-made tunes from AI mimicry? We learn:
“Lanternier said Deezer’s home-grown detection tool was able to spot markers of AI provenance with 98% accuracy. ‘An audio signal is an extremely complex bundle of information. When AI algorithms generate a new song, there are little sounds that only they make which give them away… that we’re able to spot,’ he said. ‘It’s not audible to the human ear, but it’s visible in the audio signal.’"
Will bots find a way to eliminate such tell-tale artifacts? Will legislation ever catch up to reality? Will Big Streaming feel pressure to implement their own measures? This will be an interesting process to follow.
Cynthia Murrell, July 14, 2025
Just What You Want: Information about Footnotes
July 11, 2025
No smart software to write this essay. This dinobaby is somewhat old fashioned.
I am completing my 14th monograph. Some of these 150 page plus documents became books. Examples range from The Google Legacy, published in 2003 for a client and then as a public document in 2004 by Infonortics Ltd., a specialty publisher somewhere in England. Others were published by Panda Press in Sweden. Martin White and I published a book about enterprise search management, and I do not recall what outfit published the book. When I started writing texts to accompany my lectures for ISS Telestrategies, the US National Cyber Crime events, and other specialized conferences, I decided to generate Adobe PDF files and make these “books” available to those in my classes and lectures. Dark Web Notebook and CyberOSINT were “self published.” Why? The commercial specialty publishers were going out of business or did not have a way to market the books I wrote. I wrote a couple of monographs about Japan’s investments in database technology in the early 1990s for the US Office of Technology Assessment. But I have lost track of these “books.”
When I read “Give Footnotes the Boot,” I thought about how I had handled “notes” in my long form writings. For this blog which is a collection of “notes” to myself given the appearance of an essay, I usually cite an article. I then add my preliminary thoughts about the write up, usually including a couple of the source document’s “interesting” statements. The blog, therefore, is an online notebook with 20,000 plus entries written for an audience of one: Me.
I noted that the cited “footnote” article says:
If the footnote markers are links, then the user can use the back button/gesture to return to the main content. But, even though this restores the previous scroll position, the user is still left with the challenge of finding their previous place in a wall of text6. We could try to solve that problem by dynamically pulling the content from the footnotes and displaying it in a popover. In some browsers (including yours) that will display like a tooltip, pointing directly back to the footnote marker. Thanks to modern web features, this can be done entirely without JavaScript7. But this is still shit! I see good, smart people, who’d always avoid using “click here” as link text, littering their articles with link texts such as 1, 7, and sometimes even 12. Not only is this as contextless as “click here”, it also provides the extra frustration of a tiny-weeny hit target. Update: Adrian Roselli pointed out that there are numerous bugs with accessibility tooling and superscript. And all this for what? To cargo-cult academia? Stop it! Stop it now! Footnotes are a shitty hack built on the limitations of printed media. It’s dumb to build on top of those limitations when they don’t exist on the web platform. So I ask you to break free of footnotes and do something better.
The essay omits one option; that is, just write as if the information in the chapter, book, paragraph is common knowledge. The result is fewer footnotes.
I am giving this footnote free approach a try in the book I am working on to accompany my lectures about Telegram for law enforcement, cyber attorneys, and intelligence professionals. I know that most people do not know that a specific quote I include from Pavel Durov originated from a Russia language blog. However, citing the Russian blog, presenting the title of the blog post in Cyrillic, including the English translation, and adding comments like “no longer online” would be the appropriate way to let my reader know I did not make up Pavel’s statement about having more than 100 children.
I am assuming that every person on earth knows that Pavel thinks he is a super human and has the duty to spawn more Pavels.
How will this work out? My hunch is that my readers will use my Telegram Labyrinth monograph to get oriented to a service alleged to be a criminal enterprise by the French judiciary. If someone wants to know where one of my “facts” originates, I will go through my notes, including blog posts, for the link to the document I read. Will those sources be findable in 2025 when the book comes out? Probably not.
Online information is disappearing at an alarming rate. The search systems I use “disappear” content even though I have a PDF of the source document in my electronic file. Intermediaries go out of business or filters block access to content.
I like the ideas in Jake Archibald’s essay. I also like the academic rigor of footnotes. But for the Telegram Labyrinth, I am minimizing footnotes. I assume that every investigator, intelligence professional, and government lawyer will know about Telegram. Therefore, what’s in my new book is common knowledge. That means, “Sorry, Miss Dalton, Stevie is dumping 95 percent of the footnotes.” (I should footnote that Miss Dalton was one of my teachers who wanted footnotes in Modern Language Association style for everything single thing her students wrote.) Nope. Blame Web rot, blame my laziness, blame the wild social media environment.
You will live and probably have some of that Telegram common knowledge refreshed; for example, the Telegram programming language FIFT is like FORTH only better. Get the pun. The Durovs have a sense of humor.
Stephen E Arnold, July1, 2025
Google and the EU: A Couple That Do Not Get Along
July 11, 2025
Google’s EU legal woes are in the news again. The Mercury News shares the Bloomberg piece, “Google Suffers Setback in Fight Over EU’s 4.1 Billion Euros Fine.” An advisor to the EU’s Court of Justice, Advocate General Juliane Kokott, agrees with regulators’ choice to punish google for abusing Android’s market power and discredits the company’s legal arguments. She emphasized:
“Google held a dominant position in several markets of the Android ecosystem and thus benefited from network effects that enabled it to ensure that users used Google Search. As a result, Google obtained access to data that enabled it in turn to improve its service.”
Though Kokott’s opinion is not binding, the court is known to rely heavily on its adviser’s opinions in final rulings. For its part, Google insists any market advantage it has is solely “due to innovation.” Sure, rigging the Search environment in its favor was plenty innovative. Just not legal. Not in the EU, anyway. Samuel Stolton reports:
“The top EU court’s final decision could prove pivotal for the future of the Android business model — which has provided free software in exchange for conditions imposed on mobile phone manufacturers. Such contracts provoked the ire of the commission in 2018, when the watchdog accused Alphabet Inc.’s Google of three separate types of illegal behavior that helped cement the dominance of its search engine, accompanying the order with the record fine. First, it said Google was illegally forcing handset makers to pre-install the Google Search app and the Chrome browser as a condition for licensing its Play Store — the marketplace for Android apps. Second, the EU said Google made payments to some large manufacturers and operators on condition that they exclusively pre-installed the Google Search app. Lastly, the EU said the Mountain View, California-based company prevented manufacturers wishing to pre-install apps from running alternative versions of Android not approved by Google.”
Meanwhile, the company is also in hot water over the EU’s Digital Markets Act. We learn that, in March, regulators scolded the firm elevating its own services over others and actively preventing app developers from guiding users to offers outside its app store. These practices violate the act, Google was told, and continuing to do so could lead to more fines. But are fines, even $4 billion ones, enough to deter the tech giant?
Cynthia Murrell, July 11, 2025
Win Big at the Stock Market: AI Can Predict What Humans Will Do
July 10, 2025
No smart software to write this essay. This dinobaby is somewhat old fashioned.
AI is hot. Click bait is hotter. And the hottest is AI figuring out what humans will do “next.” Think stock picking. Think pitching a company “known” to buy what you are selling. The applications of predictive smart software make intelligence professionals gaming the moves of an adversary quiver with joy.
“New Mind-Reading’ AI Predicts What Humans Will Do Next, And It’s Shockingly Accurate” explains:
Researchers have developed an AI called Centaur that accurately predicts human behavior across virtually any psychological experiment. It even outperforms the specialized computer models scientists have been using for decades. Trained on data from more than 60,000 people making over 10 million decisions, Centaur captures the underlying patterns of how we think, learn, and make choices.
Since I believe everything I read on the Internet, smart software definitely can pull off this trick.
How does this work?
Rather than building from scratch, researchers took Meta’s Llama 3.1 language model (the same type powering ChatGPT) and gave it specialized training on human behavior. They used a technique that allows them to modify only a tiny fraction of the AI’s programming while keeping most of it unchanged. The entire training process took only five days on a high-end computer processor.
Hmmm. The Zuck’s smart software. Isn’t Meta in the midst of playing catch up. The company is believed to be hiring OpenAI professionals and other wizards who can convert the “also in the race” to “winner” more quickly than one can say “billions of dollar spent on virtual reality.”
The write up does not just predict what a humanoid or a dinobaby will do. The write up reports:
n a surprising discovery, Centaur’s internal workings had become more aligned with human brain activity, even though it was never explicitly trained to match neural data. When researchers compared the AI’s internal states to brain scans of people performing the same tasks, they found stronger correlations than with the original, untrained model. Learning to predict human behavior apparently forced the AI to develop internal representations that mirror how our brains actually process information. The AI essentially reverse-engineered aspects of human cognition just by studying our choices. The team also demonstrated how Centaur could accelerate scientific discovery.
I am sold. Imagine. These researchers will be able to make profitable investments, know when to take an alternate path to a popular tourist attraction, and discover a drug that will cure male pattern baldness. Amazing.
My hunch is that predictive analytics hooked up to a semi-hallucinating large language model can produce outputs. Will these predict human behavior? Absolutely. Did the Centaur system predict that I would believe this? Absolutely. Was it hallucinating? Yep, poor Centaur.
Stephen E Arnold, July 10, 2025
BBC Warns Perplexity That the Beeb Lawyers Are Not Happy
July 10, 2025
The BBC has had enough of Perplexity AI gobbling up and spitting out its content. Sometimes with errors. The news site declares, “BBC Threatened AI Firm with Legal Action over Unauthorised Content Use.” Well, less a threat and more a strongly worded letter. Tech reporter Liv McMahon writes:
“The BBC is threatening to take legal action against an artificial intelligence (AI) firm whose chatbot the corporation says is reproducing BBC content ‘verbatim’ without its permission. The BBC has written to Perplexity, which is based in the US, demanding it immediately stops using BBC content, deletes any it holds, and proposes financial compensation for the material it has already used. … The BBC also cited its research published earlier this year that found four popular AI chatbots – including Perplexity AI – were inaccurately summarising news stories, including some BBC content. Pointing to findings of significant issues with representation of BBC content in some Perplexity AI responses analysed, it said such output fell short of BBC Editorial Guidelines around the provision of impartial and accurate news.”
Perplexity answered the BBC’s charges with an odd reference to a third party:
“In a statement, Perplexity said: ‘The BBC’s claims are just one more part of the overwhelming evidence that the BBC will do anything to preserve Google’s illegal monopoly.’ It did not explain what it believed the relevance of Google was to the BBC’s position, or offer any further comment.”
Huh? Of course, Perplexity is not the only AI firm facing such complaints, nor is the BBC the only publisher complaining. The Professional Publishers Association, which represents over 300 media brands, seconds the BBC’s allegations. In fact, the organization charges, Web-scraping AI platforms constantly violate UK copyrights. Though sites can attempt to block models with the Robots Exclusion Protocol (robots.txt), compliance is voluntary. Perplexity, the BBC claims, has not respected the protocol on its site. Perplexity denies that accusation.
Cynthia Murrell, July 10, 2025
Apple and Telegram: Victims of Their Strategic Hubris
July 9, 2025
No smart software to write this essay. This dinobaby is somewhat old fashioned.
What’s “strategic hubris”? I use this bound phrase to signal that an organization manifests decisions that combine big thinking with a destructive character flow. Strategy is the word I use to capture the most important ideas to get an organization to generate revenue and win in its business and political battles. Now hubris. A culture of superiority may be the weird instinct of a founder; it may be marketing jingo that people start believing; or it is jargon learned in school. When the two come together, some organizations can make expensive, often laughable, mistakes. Examples range from Windows and its mobile phone to the Ford Edsel.
I read “Apple Reaches Out to OpenAI, Anthropic to Build Out Siri technology.” In my opinion, this illustrates strategic hubris operating on two pivot points like a merry-go-round: Up and down; round and round.
The cited article states:
… over the past year or so it [Apple] has faced a variety of leadership and technological challenges developing Apple Intelligence, which is based on in-house foundation models. The more personalized Siri technology with more personalized AI-driven features is now due in 2026, according to a statement by Apple …
This “failure” is a result of strategic hubris. Apple’s leadership believed it could handle smart software. The company taught China how to be a manufacturing super power could learn and do AI. Apple’s leadership seems to have followed the marketing rule: Fire, Aim, Ready. Apple announced AI or Apple Intelligence and then failed to deliver. Then Apple reorganized and it failed again. Now Apple is looking at third party firms to provide the “intelligence” for Apple.
Personally I think smart software is good at some things and terrible at others. Nevertheless, a failure to provide or “do” smart software is the digital equivalent of having a teacher put a dunce cap on a kid’s head and making him sit in the back of the classroom. In the last 18 months, Apple has been playing fast and loose with court decisions, playing nice with China, and writing checks for assorted fines levied by courts. But the premier action has been the firm’s failure in the alleged “next big thing”.
Let me shift from Apple because there is a firm in the same boat as the king of Cupertino. Telegram has no smart software. Nikolai Durov is, according to Pavel (the task master) working on AI. However, like Apple, Telegram has been chatting up (allegedly) Elon Musk. The Grok AI system, some rumors have it, would / could / should be integrated into the Telegram platform. Telegram has the same strategic hubris I associated with Apple. (These are not the only two firms afflicted with this digital SARS variant.)
I want to identify several messages I extracted from the Apple and Telegram AI anecdotes:
- Both companies were doing other things when the smart software yachts left the docks in Half Moon Bay
- Both companies have the job of integrating another firm’s smart software into large, fast-moving companies with many moving parts, legal problems, and engineers who are definitely into “strategic hubris”
- Both companies have to deliver AI that does not alienate existing users and attract new customers at the same time.
Will these firms be able to deliver a good enough AI solution? Probably. However, both may be vulnerable to third parties who hop on a merry-go-round. There is a predictable and actually no-so-smart pony named Apple and one named Messenger. The threat is that Apple and Telegram have been transmogrified into little wooden ponies. The smart people just ride them until the time is right to jump off.
That’s one scenario for companies with strategic hubris who missed the AI yachts when they were under construction and who were not on the expensive machines when they cast off. Can the costs of strategic hubris be recovered? The stakeholders hope so.
Stephen E Arnold, July 9, 2025
-
- Subscribe to Beyond Search
Feature archive
News archive
-
