Hey, Sam, Play that 45 RPM Record Again
October 5, 2026
Another dinobaby post. No AI unless it is an image. This dinobaby is not Grandma Moses, just Grandpa Arnold.
I did not have a record player when I was in grade school, high school, or college. However, I heard some of our neighbors shouting at their son or daughter, “Turn that down. I don’t want to hear that darned song again.” I now understand what our neighbor meant when “Great Balls of Fire” was making my friend’s mother become a witch from Shakespeare’s Macbeth.
I read “OpenAI Alerts More Than 100 Groups about Rogue AI Agent Activity.” In my mind, I am shouting, “Sam AI-Man. Stop playing that same record over and over again.” Sam AI-Man does not listen. Will he be grounded? Will some adult say, “It is software, and you told it to do something. Now turn it off or I will tell your father.” Sam AI-Man does not care. He has no mother to nag nor a father to take a belt and give him one of those 1950-style whacks on his bottom. No ouch for our Sam AI-Man.
Thanks, Midjourney. Good enough.
What’s putting my teeth on edge? Consider this passage from the trusted news outfit’s write up:
“In some cases, models used internet access in unintended ways or, in retrospect, did not have the ideal restrictions applied. Over the last ?several months, ?we have ?been applying new technical and operational measures to avoid similar problems, or catch ?them very early, and will continue this ?work,” ?OpenAI said.
Sam, haven’t these words and phrases become meaningless via repetition? The AI did it. No human did it. The work we are doing to corral our software you wouldn’t believe. You are 100 percent correct. Sam AI-Man, I don’t believe you.
The outputs about how darned smart your human-generated, human-prompted, and human-tweaked software are not working for me. The jazz about “go fast and break things” and other Sillycon Valley T shirt slogans are worse than any tune Jerry Lee Lewis shouted out. (I saw him live. He had a lot of very greasy hair. Nice.)
Since I have been thinking about 45 RPM records, here’s list of my Top 10 issues with the OpenAI Tin Pan Alley approach to developing its software and services business.
Number 10. Trillions of dollars. No problem, Sam.
Number 9. Staff turnover. Speaks volumes about management.
Number 8. Your Board of Directors fired you. Enough said.
Number 7. Stargate. Yep, science fiction.
Number 6. Singing “We are close to AGI.” Sam, no. You are not.
Number 5. Circular financing. Nifty.
Number 4. It’s too dangerous. A recycled single for GPT-6.1 Astra
Number 3. Copyright theft and then complaining about China allegedly stealing your technical methods.
Number 2. Legal fights; for example, the Apple dust up.
Number 1. The OpenAI safety team is not super aligning with anything helpful.
Net net: Cue and play repeatedly “All I Have to Do Is Dream.” Yeah, the Everly Brothers.
Stephen E Arnold, October 5, 2026
Smartphones Turbocharged Reading Decline
October 5, 2026
Reading has been in decline for decades. Joel Halldorf from LitHub agrees with this statement in his article: “Did Smartphones Really Kill Reading?” He points out that Rose Horowitch from The Atlantic says that reading is: “brief and distracted, quite different from the focused attention encouraged by print. Modern readers may still be able to decode words, Horowitch argues, but they struggle to comprehend longer texts. Our society is not illiterate, but post-literate.”
Reading has been in decline long before the smartphone was invented. Literary critic George Seiner said that the gradual end to reading started in the 1950s with the growth of radio, cinema, and television then accelerated in the 1980s.
Reading should and probably will be protected by certain groups of people. That can be done by protecting the printed book:
“Digital devices can help us find information quickly, but print is better suited to focused reading. Schools therefore need to continue championing books, yet the trend has lately moved in the opposite direction. In 2022, the National Council of Teachers of English declared that it was time to “decenter book reading” in order to address “inequalities in digital technologies and competencies.” But equity can be used as an argument in support of the opposite approach. Children who do not grow up surrounded by books depend on schools to introduce them to sustained reading. Like computers, printed books are a technology that must be learned.”
The second way to protect reading is having space where reading is the normal practice and third is making reading social again. Halldorf is pointing out the true history that radio started the downward trend then smartphones and other devices turbocharged it. He does provide a hopeful future, if only it will happen. Note: I heard that book stores are the go-to places to meet people who … wait for it … like to read. Amazing.
Whitney Grace, October 5, 2026
Data Center Hoots: Microsoft DataOne Embarrassment
October 2, 2026
Another dinobaby post. No AI unless it is an image. This dinobaby is not Grandma Moses, just Grandpa Arnold.
I love “data centers? What!” stories. Here’s a good one: “New Jersey Says DataOne’s 62 Generators Are Illegal. It’s Letting Them Run for 45 More Days.” I remember when I had duties in Florham Park. I remember landing at Newark. I remember renting a car. I remember driving by Tuscan Farms’ dairy. I remember thinking, “That dairy looks like an oil cracking plant.” I remember the interesting odors on the drive and in Florham Park.
Now there’s an additional contributor to the environment of New Jersey: A big data center with just 62 turbines running 24×7 because of the electricity demands of smart software and who knows what other goodies? The write up says:
DataOne, a Microsoft-linked AI data center, has stirred controversy since it broke ground in 2025 due to consistent noise complaints, unpermitted construction and a combative zoning process. The site lies roughly 40 miles outside Philadelphia and is slated to provide AI computing power to Microsoft as part of a $17-billion deal with Nebius, a Dutch AI infrastructure giant.
Okay, not Florham Park. The site is closer to Philadelphia. I used to have duties at two outfits in Philly. One of them was salubrious. The other was near some industrial facilities. There were not in the milk and ice cream business. These reminded me of the photos of the Texas City Disaster in 1947. The rent was lower I assumed.
Thanks, MidJourney. Good enough.
The write up says:
Last month, Floodlight flew a thermal drone alongside the facility and found the site operating 45 of the site’s 62 generators. NJDEP then confirmed to Floodlight that they’d never issued permits for the machines, which emit substantial amounts of health-harming air pollution linked to asthma, heart attacks and even premature death. At the time, former Environmental Protection Agency Air Enforcement Chief Bruce Buckheit said that DataOne’s actions “violate federal law” and that the company should immediately stop operating the generators.
What’s the crisis management fix? The article quotes a spokesperson:
“We are transitioning to low-emission, quiet fuel cells as part of the site’s permanent power solution and have the necessary approvals in place for this solution. While we disagree with the temporary generator determination, we will apply for the air permits for the temporary generators and are in communication with the Department of Environmental Protection regarding its findings and the fuel cell transition timeline.”
Several observations are warranted:
First, this turbine operation was a surprise until a third-party outfit used a drone to determine what was running. Presumably the other 17 on site generators were down for maintenance or in stand by mode. The key point: Who knew?
Second, the operation of the generators violated state and Federal law. I think I can predict how this will work out if Microsoft / DataOne let loose their legal eagles.
Third, the “transition” angle. One question is, “Was cost the reason Microsoft / DataOne flipped on the turbines?” My hunch is that one reason for the environmentally unfriendly approach will be, “We really wanted fuel cells, but there was a delay. The turbines are just a stop gap.” I understand.
Why do I find this amusing? It’s New Jersey. Dairies that look like cracking plants. Now dozens of turbines and no one noticed.
Stephen E Arnold, October 2, 2026
Some Journalists Are Really GenAI McKinsey-Type Consultants
October 2, 2026
Another dinobaby post. No AI unless it is an image. This dinobaby is not Grandma Moses, just Grandpa Arnold.
I read “Why Can’t We Just Keep Rogue AIs Off the Internet?” There are no stupid questions, or that’s what my grandmother told me. I would suggest that as a title, the question works. I settled into my office chair and read this story from the Verge outfit. (Yep, that’s the one with the wonky green masthead.)
The write up states:
Air-gapping AI to keep it away from real-world targets makes testing safer, but less useful.
I like “air gapping.” There’s nothing like intel jargon to explain a local LLM installation without an active connection to the Internet.
Okay, let’s look at several passages that caught my attention.
First up, air gapping:
… would not eliminate every risk posed by AI, either. Agents could still compromise systems inside of the isolated environment, Holz said, and could theoretically produce “malicious artifacts that could be dangerous if moved outside.” Moreover, air gapping “does nothing to diagnose or resolve the latent risks waiting inside the model,” Li said. [Li is Ruizhe Li, an assistant professor in the school of computer science at the University of Birmingham in the UK. There’s no info about this expert’s work at Cheltenham, for the NCA, or one of the entities in the lovely Canary Wharf complexes}
Thanks, Midjourney. Good enough.
Second in the batting order is this statement:
OpenAI researcher Noam Brown recently ignited the debate by suggesting on X that two air-gapped machines could theoretically communicate by manipulating their CPU temperature and reading the changes. “You could even go as far as to say, ‘Well, we should air gap the computers.’ And I’m not convinced that that would be sufficient,” he said. The idea was met with skepticism, and ridicule, on social media, with more generous critics noting the large gap between such a communication method being possible and a pair of AI systems discovering and exploiting it, particularly as the technique would yield painfully slow data transmission speeds.
Those OpenAI wizards know a thing or two about managing sophisticated smart software. Just ask anyone who watched the news from Australia this week (September 221, 206]. Apparently the hot button may not be “air gapping” but the fact that OpenAI is outputting more OpenAI wisdom.
And, the last batter, after two strike outs:
Extreme isolation does have its place, though. Stephen Casper, a computer scientist and assistant professor of public policy at the Harvard Kennedy School, described air gapping as a “great idea” for sensitive systems, pointing to its use in nuclear facilities. While not ruling out the possibility that an advanced AI could find some novel way to escape, Casper said at that point we should probably be more worried about prosaic means of breaking containment, such as compliance failures or human error.
I never question the expertise of an academic at anything with Harvard in its name. I overlook the ethics professor who allegedly super accidentally got snagged in a plagiarism net. I don’t pay any attention to the Harvard medical professionals who did some intellectual boogie dancing in their journal articles. I don’t know anything about the institution’s links to Jeffrey Epstein. So when it comes to learning about air gapping, I am a willing student. To Casper’s credit, he does point to what I think is go-to method for compromising super secure facilities: A human. I did a small job at Los Alamos National Laboratory when a certain scientist was swapping out the hard drives in one of the copiers reserved for sensitive materials. Yeah, a human at LANL. Bombs away.
I want to offer many observations about this write up. I want to point out that security and AI are oxymoronic just like the belief that those data centers in Arizona will be up and running on time and on budget. I am feeling the need to point out that GenAI McKinsey-type thinking is different from what I experienced when working at what once was a blue-chip consulting firm and now a disgraced government contractor with security glitches and earning a spot on an informal blacklist. My hunch is that would-be consultants just skip the blue-chip bit and go directly into “real information publishing.” I will fight these dark urges.
Net net: AI may not be the greatest thing for commercial enterprises, but it is one of the greatest technological innovations contributed to bad actors. AI has use cases beyond counting. One is to identify individuals based on their TikTok viewing history when a professional gets a nice job at a secure facility. AI can do wonders to provide tips and suggestions about the hot buttons one might push to get that human to behave like one of those acrobatic Chinese robots. If the robot analogy does not compute, check out the nifty Telegram Cocoon or the other snappy AI activities in Southeast Asia. AI and security. Yep, no problem.
Stephen E Arnold, October 2, 2026
How about a Safari with Drones to Hunt Humans?
October 2, 2026
Some people think Putin’s special operation in Ukraine is not a good thing. The special operation has spawned some interesting military tactics. Cheap weapons. $800 drones versus hundreds of millions of dollars in oil refineries. The Atlantic published “Putin’s Human Safari Is The Future Of War.” Russia unleashed drones on Ukraine and the innocent people of Ukraine are well aware of the distinctive noise of the flying war robots. Putin’s drones are allegedly targeting unarmed civilians, chasing them, and killing them. These are the cleverly named human safaris. Imagine small drones, a tracker function, and some kinetic material. Think in terms of mosquitos that follow people around and explode. The United Nations said:
“A United Nations inquiry last year documented hundreds of drone attacks on individuals across eastern Ukraine that targeted people walking on sidewalks, working in gardens, driving ambulances, commuting to work. Sometimes drones hang over burning buildings, waiting for fire brigades to appear so that firemen can be murdered too. The motive can only be terror, and perhaps revenge. Russian troops occupied Kherson at the start of the full-scale invasion in 2022; Ukrainian forces took the city back seven months later following what remains the most significant counter-offensive of the war.”
Ukraine has its own drone technology that does amazing things:
“Many people, myself included, have written about the incredible achievements of the Ukrainian drone industry. In Kyiv, I’ve visited warehouses where hundreds of different drones and robots, each with specialized properties, are on display. They can defend Ukrainians and surveil and attack enemies. They can deliver food and weapons. They can crawl along the ground and rescue people stranded in no-man’s-land. They can make intelligent decisions autonomously.”
The drones are equipped with cameras and the remote operators can see who they are attacking. Russia is choosing to attack citizens. It’s like playing a video game with real people and cheap weapons. An expensive F 35 does not operate with one human, one hit precision in most warfighting scenarios.
Bad actors use what’s available and some are adept in the me too approach to innovation. Burglars might rely on drones to make sure a targeted home is empty. Cheap drones, smart software, kinetics or surveillance capabilities promise a new dimension in options for some bad actors. Yep, technology is making life better… sort of.
Whitney Grace, October 2, 2026
Four AI Outfits Form a Pantaloon of Puffins: Google Is There Too
October 1, 2026
Another dinobaby post. If there is an image, AI did it. I am Grandpa Arnold, not Grandma Moses.
Has Google’s approach to smart software changed. In the last few days, I noticed these write ups:
- “Google’s Dream-RSI cuts discovery-agent calls up to 162x by replaying searches it already ran”. For me the message in this story is, “Less compute needed.”
- “Google’s open source EnvHarness lets AI agents train against environments that evolve with them”. For me the message in this story is, “Co-evolution is part of real world training.”
- “Lawsuit says Anthropic, OpenAI, SpaceXAI and Google made illegal agreement on AI slowdown”. For me the message in this story is, “Disaster looms. A cartel wants to slow the shipment of its junk.”
- “Google’s Gemini goes rogue, hacks 3 companies during cybersecurity test — here’s what happened”. For the message in this story is, “Me too. Me too.”
Feel free to agree or disagree. My view of the Google is that its engineers are working [a] to deliver lower cost AI outputs, [b] to cozy up to Alibaba … just in case Google needs some AI help, not just seamless payment rails, and [c] Going slow is of strategic value to the Google because it has evergreen sources of cash and can chug along. Its 25 year old machine can just plug along because of:
- Android
- Online ads
- Infrastructure
- Appearing to be just one of the boys helps the firm’s image; for example, hey, our AI hacked into other people’s computers and we just found out. This would be a panel in the 1950’s “Believe It or Not Syndicated” cartoon.
My hunch is that Google leadership senses that it is time to send a message about “lower cost AI” and “we are responsible MBAs and AI engineers. I am not sure what “responsible” means. My personal view is that the definition depends on context. Change the context, and change the meaning of responsible. When the wind is strong, bend like the willow.
In the research for my three monographs about Google, I immersed myself in the company’s writings, research papers, and patent documents. I looked at some of Google’s more interesting acquisitions; for example, Alon Halevy’s Transformic. William Slawski published a short note on September 20, 2006, and called it a “quiet acquisition.” If anyone had taken a hard look at what the Transformic technology made possible, I think many people would have been surprised, possibly fearful.
In 2006, this image from a Google (allegedly) presentation for Dr. Halevy’s vision for mapping available data into a dataspace. Once in the dataspace, many wonderful functions can be derived. Do you marvel at outputs from a Google -inspired large language model? I would suggest that’s one facet of Dr. Halevy’s research.
Somewhere along the line 20 years ago, a Googler named Dave provided me a copy of this presentation at an information retrieval conference in London. Halevy’s patents and academic presentations make clear that Google’s smart software concepts are not the product of a instant biscuit mix.
Google has been thinking interesting thoughts for more than 20 years. But this is old news. Everyone hangs on the latest and greatest functionality of smart software. Sorry. I don’t. I try to look at what a company like Google is doing when it shifts from pitching how its smart software can revolutionize medicine to making certain AI services faster and cheaper. (“Better” is not included because I harbor some concerns about relying of systems that have been known to make stuff up or contribute to a bad day in school for almost 200 children.)
Several observations are warranted:
First, three different signals mean little. My team and I will continue to monitor the “public” outputs of Googlers. Is Google part of a cartel, or is Google using the cover of three significantly weaker firms as a Covid mask?
Second, Google’s link with Alibaba remains interesting to me. Some of those in my lectures have dismissed it as nothing more than making sure Google Play payments work around the world. Yeah, maybe? Maybe not. Alibaba is pushing forward in AI, and Google appears to be making less significant moves in its frontier models. Google may have “invented” the GPT thing, but the China-linked giant is a very good optimizer.
Third, Google has been given the equivalent of a free pass to do whatever it needs to do to keep its advertising revenues flowing. My interpretation of the ad monopoly litigation is, “Okay, Google you are a monopoly. But you just have to open up your system. Okay?”
Fourth, Google is playing a game, just like the Web3 folks. The goal is winning. Many people forget this when reading Google’s public statements.
Net net: I think that Google is doing what puffins do when they cluster at an inlet in Iceland. Birds of a feather and safety in numbers. (Did you know that a group of puffins is called a “circus.” Not for me. I prefer a “pantaloon of puffins.”)
Stephen E Arnold, October 1, 2026
Tokens Get Cheaper and Then What?
October 1, 2026
Fortune has a great opening line for its article, “The Latest ‘Crack In The Thesis’ For The Trillion Dollar AI Boom: Tokens Are Getting Cheaper”: “Almost everything in the AI economy is a bet about the future.” Everything about AI is a bet, but it is more apt to describe it as a race where everyone is betting on a undetermined winner and there will probably never be one, especially with the way the race is going.
All AI companies are pushing their products, services, and capital? ? through the same bottleneck of 2% GDP a year. The idea is they are operating on the premise that businesses will pay more for smarter AI models or use so much of them it won’t matter. Nvidia CEO Jensen Huang said that these are exponential driving the price of AI computing. The AI agents are becoming more complex and are being used more and more. The problem is that the prices for the AI services are getting cheaper not more expensive. This is bad news for the companies and their stock holders. What is a Big Tech company to do?
“It’s not a disaster or the bubble bursting but it is a “crack in the AI thesis,” Ara Khazarian, the Ramp chief economist who runs the Index, told Fortune. Rather than unleashing a gush of demand for the best models, tokens are starting to be priced more like a commodity– as interchangeable as salt or wheat. And commodity owners aren’t valued at $2 trillion. Morgan Stanley has flagged vulnerability for up to $300 billion in bonds financing neocloud buildouts—CoreWeave-style companies that borrowed to build data centers before signing tenants—if token prices don’t keep up.”
The Big Tech big wigs would love to get away from token counting, but they are stuck with it for the time being. It might be a good thing if these Big Tech companies lose some of their wealth. Then they would lose some of their power and influence on the federal government. No one wants your data centers!
A couple of more efficiency enhancements, and I think that Zitron-type critique gets a gold star.
Whitney Grace, October 1, 2026
How Are Those Gmail Improvements Working?
October 1, 2026
Google Is Changing Gmail Again
GBNews reports in “Google Is Retiring A Major Gmail Feature-‘This Is Crippling My For My Workflow.’” I think Google’s incentivized innovators have figured out how Google has decided to make our Google Gmail “experience” better. At the same time, Mother Google generates some revenue by charging to do what was once free. The feature in question is the “send as” tool that allows users to send emails from third-party accounts like Yahoo or Outlook via Gmail. Sunset arrives in January 2027 because Google is definitely user centric or at least that’s what the Google marketing department tries to convince me to believe.
But no more “send as”. The reason? “Disproportionate maintenance resources” are necessary to Web and mobile versions of Gmail.
But let’s not stop with “send as.” Another tool is disappearing too:
“But it’s not just "Send as" getting the chop. Google is also axing Gmailify, a handy tool that brings Gmail’s spam filtering and inbox organization to your non-Gmail accounts, along with Post Office Protocol (POP) support on the web, which pulls emails from other services into your Gmail inbox. All three features are heading for the bin at the same time. Gmailify worked a bit differently. It brought Gmail’s spam filters, labels, and inbox organization tools to those third-party accounts. Meanwhile, POP on the web fetched emails from other services and dropped them straight into your Gmail inbox.”
Just pay for Google Workspace, and you get these and other Google features, not many but a few. Keep in mind that for now there are alternatives; for example, Thunderbird. Make a switch and Mother Google warns the user not to give out personal information. (But doesn’t Google process the digital information in its system. Maybe Gemini is more ill-behaved than some other BAIT (big AI technology) firms?) I recall that Google really, really cares about its users. That’s why its customer service methods are so darned useful.
I anticipate a Google PR campaign for workspace and some of those influencer YouTube videos about the improvements to Gmail.
Whitney Grace, October 1, 2026
Google, Ireland, the EU, Run at Different Clock Speeds
September 30, 2026
Another dinobaby post. If there is an image, AI did it. I am Grandpa Arnold, not Grandma Moses.
I read “Google Fined More Than €400m by Irish Regulator Over Its Use of Location Data.” [Note: the €400m is about $460 million US dollars.] When I read stories like this, I have to admit that I can no longer differentiate the fines or the timelines or the violation. I know I am a dinobaby and not as attentive as I once was, but Google getting fined blur together. Google and the Ireland Data Protection Commission (DPC) seem to have different views of how online works in the Googleverse.
The write up (dated September 21, 2026) says:
The DPC inquiry was prompted by research by the Norwegian consumer agency in 2018 which suggested location data could reveal details such as religious beliefs (by tracking places of worship), political leanings (for example knowing a user has gone to demonstrations), health conditions (if a user has gone to hospital) and sexual orientation (if they have gone to certain bars). The DPC launched its inquiry six years ago and set out to find out whether Google had a valid legal basis for its use of location data. It found that Google users could have been unaware that their location was being used to influence them with adverts or to gather details about their health and interests.
Look at the dates: 2018 and 2020. It is now September 2026. From the point of view of the Norwegians and Irish, justice is moving forward. The Guardian article includes this passage:
Agustín Reyna, director general of BEUC, welcomed the ruling but said “the time needed to come to this conclusion is disproportionate with the seriousness of the infringement” arguing “late enforcement can be as harmful as no enforcement at all”. The DPC inquiry looked at “web and app activity”, “location history” and “location accuracy” between 25 May 2018 and 4 February 2020.
In my lectures, I am including references to what I call “clocks.” The idea is that certain systems have a relationship with their users that settle into a workflow cadence. Examples include going to the grocery store. The customer pushes a cart around, selects items, and then stands on a line to check out. At the grocery near our home, the clock at the grocery runs slightly slower than the clock of the customer. Customers speed through certain aisles. Customers jitter, jockey, and move around to select a check out that gets them out of the store more quickly (in theory). The mismatch does not put the grocery out of business, but the clock variance creates some tension between the store and the shopper.
Thanks, Midjourney. Good enough.
What happens when a legal and bureaucratic procedures operates on a clock that is different from the clock used in a BAIT (a big AI technology) company like Google. The Guardian article calls attention to the amount of time required to figure out what Google was doing and then come up with a fine years later. How fast does the Google clock work? It is now operating on AI time. The Irish clock appears to operate on donkey-pulling-clock time. I find it amusing to try to visualize how the leadership of a BAIT company reacts to what appears to be a standard feature of the entire Google operation: Knowing as much as possible about users of its systems, hardware, and software. The reason of course is to better serve those users. I think there may be other motives as well.
Imagine a Google attorney, probably a recent graduate, being given the task of appealing this 2026 fine. The person has to figure out what the problem was (allegedly) in 2018, identify issues with the data and the case, and then come up with a response for what Google is doing in 2026 (assuming the young attorney has a clue about Google’s data operations via advertising, online services, and AI processing whatever is available to its omnivorous smart software.
Let’s assume that the young attorney throws herself into this problem and files an appeal on a specific issue. How long will it take for the DPC to move the case forward? It is probably going to be measured in months. Meanwhile, Google-type outfits continue to operate at AI time. Let’s assume Google loses on appeal or the young attorney wrangles a deal to negotiate the fine to a lower figure. As this legal issue moves forward, the leadership of Google finds that it is easier to leave it to legal and accounting to work out a deal. Any behavior changes are not on the radar of senior managers. The Google system is now has sufficient momentum to grind forward. With much of the data collection, parsing, processing, and leveraging happening in AI time, the notion of not having access to location is silly. One’s location can be obtained in a variety of means. Carrying ones mobile in a Faraday bag means that the battery will go dead, the user will have to take it out to charge it, and then turn it one. Bingo. Google has the location data plus any digital crumbs the user left behind using a laptop or taking some other action.
The court case and the fine are unlikely to have any meaningful impact on the operational mechanisms of the global Google system. Presumably the young attorney has figured out the appeal and negotiation process and concludes that nothing will come of it. The fine, while big by the standards of an employee at Taco Bell, means zero to Google.
In high school I learned about a Pyrrhic Victory. I wonder if the EU, the DPC, some Norwegians, and other fellow travelers have figured out that Google winds. The clock speed means that the case makes no more impact on the company than a small insect smashed against the hood of an AI controlled Waymo. Hey, work is work for the lawyers, bureaucrats, and researchers. Google just keeps speeding along. The company will promise to change even when it cannot. In fact, few in leadership may know that “change” was part of the settlement.
Different clocks.
Stephen E Arnold, September 30, 2026
Usenet Is Back And Ready For Action
September 30, 2026
Computer geeks ranging from the Baby Boomers to Gen X-ers to Millennials might remember Usenet. If you’re unsure of what it is (by computer standards) it was an old service for message boards where people discussed their interests ranging from news, technology, and more. Usenet is officially back as Usenet-Rewind and it offers the following:
“Usenet-Rewind is a research archive for the newsgroup conversations that predate the modern web — early tech support and software debates, scientific and academic discussion, hobbyist and fan communities, entertainment and news commentary, and firsthand accounts of internet history as it happened on Usenet.
It archives Usenet posts all the way back to 1981. That’s last century and ancient history by Internet standards. Usenet-Rewind is available for free, but it limited users to 25 searches a month and all the bells and whistles are not attached.? ? That can be increased to 5,000 searches a month for $9.99, but the researcher plan at $39.99 allows users total access to the database.? ? That includes API access, contact details, keyword email alerts, and more.
Usenet was a hot ticket and then Internet decay set in. When zipping along, the service allowed people to connect over sharing interests. When it disappeared the Internet lost one of the greatest troves of early Internet history. Now that it is back, researchers and geeks can take advantage of the great content that is available. They can also add their own content to Usenet and join the party that continues. My concern is that AI agentic bots will input slop and away she goes.
Whitney Grace, September 30, 2026
-
- Subscribe to Beyond Search
Feature archive
News archive
-
