Titanic AI Goes Round and Round: Are You Dizzy Yet?
September 23, 2025
This essay is the work of a dumb dinobaby. No smart software required.
I read “Nvidia to Invest Up to $100 Billion in OpenAI, Linking Two Artificial Intelligence Titans.” The headline makes an important point. The words “big” and “huge” are not sufficiently monumental. Now we have “titans." As you may know, a “titan” is a person of great power. I will leave out the Greek mythology. I do want to point out that “titans” were the kiddies produced by Uranus and Gaea. Titans were big dogs until Zeus and a few other Olympian gods forced them to live in what is now Newark, New Jersey.
An AI-generated diagram of a simple circular deal. Regulators and and IRS professionals enjoy challenges. What are those people doing to make the process work? Thanks, MidJourney.com. Good enough.
The write up from the outfit that it is into trust explains how two “titans” are now intertwined. No, I won’t bring up the issue of incestuous behavior. Let’s stick to the “real” news story:
Nvidia will invest up to $100 billion in OpenAI and supply it with data center chips… Nvidia will start investing in OpenAI for non-voting shares once the deal is finalized, then OpenAI can use the cash to buy Nvidia’s chips.
I am not a finance, tax, or money wizard. On the surface, it seems to me that I loan a person some money and then that person gives me the money back in exchange for products and services. I may have this wrong, but I thought a similar arrangement landed one of the once-famous enterprise search companies in a world of hurt and a member of the firm’s leadership in prison.
Reuters includes this statement:
Analysts said the deal was positive for Nvidia but also voiced concerns about whether some of Nvidia’s investment dollars might be coming back to it in the form of chip purchases. "On the one hand this helps OpenAI deliver on what are some very aspirational goals for compute infrastructure, and helps Nvidia ensure that that stuff gets built. On the other hand the ‘circular’ concerns have been raised in the past, and this will fuel them further," said Bernstein analyst Stacy Rasgon.
“Circular” — That’s an interesting word. Some of the financial transaction my team and I examined during our Telegram (the messaging outfit) research used similar methods. One of the organizations apparently aware of “circular” transactions was Huione Guarantee. No big deal, but the company has been in legal hot water for some of its circular functions. Will OpenAI and Nvidia experience similar problems? I don’t know, but the circular thing means that money goes round and round. In circular transactions, at each touch point magical number things can occur. Money deals are rarely hallucinatory like AI outputs and semiconductor marketing.
What’s this mean to companies eager to compete in smart software and Fancy Dan chips? In my opinion, I hear my inner voice saying, “You may be behind a great big circular curve. Better luck next time.”
Stephen E Arnold, September 23, 2025
AI Poker: China Has Three Aces. Google, Your Play
September 19, 2025
No smart software involved. Just a dinobaby’s work.
TV poker seems to be a thing on free or low cost US television streams. A group of people squint, sigh, and fiddle as each tries to win the big pile of cash. Another poker game is underway in the “next big thing” of smart software or AI.
Google released the Nano Banana image generator. Social media hummed. Okay, that looks like a winning hand. But another player dropped some coin on the table, squinted at the Google, and smirked just a tiny bit.
“ByteDance Unveils New AI Image Model to Rival DeepMind’s Nano Banana” explains the poker play this way:
TikTok-owner ByteDance has launched its latest image generation artificial intelligence tool Seedream 4.0, which it said surpasses Google DeepMind’s viral “Nano Banana” AI image editor across several key indicators.
Now the cute jargon may make the poker hand friendly, there is menace behind the terminology. The write up states:
ByteDance claims that Seedream 4.0 beat Gemini 2.5 Flash Image for image generation and editing on its internal evaluation benchmark MagicBench, with stronger performance in prompt adherence, alignment and aesthetics.
Okay, prompt adherence, alignment (what the heck is that?), and aesthetics. That’s three aces right.
Who has the cost advantage? The write up says:
On Fal.ai, a global generative media hosting platform, Seedream 4.0 costs US$0.03 per generated image, while Gemini 2.5 Flash Image is priced at US$0.039.
I thought in poker one raised the stakes. Well, in AI poker one lowers the price in order to raise the stakes. These players are betting the money burned in the AI furnace will be “won” as the game progresses. Will AI poker turn up on the US free TV services? Probably. Burning cash makes for wonderful viewing, especially for those who are writing the checks.
What’s China’s view of this type of gambling? The write up says:
The state has signaled its support for AI-generated content by recognizing their copyright in late 2023, but has also recently introduced mandatory labelling of such content.
The game is not over. (Am I the only person who thinks that the name “nana banana” would have been better than “nano banana”?)
Stephen E Arnold, September 19, 2025
AI: The Tool for Humanity. Do Not Laugh.
September 19, 2025
Both sides of the news media are lamenting that AI is automating jobs and putting humans out of work. Conservative and liberals remain separated on how and why AI is “stealing” jobs, but the fear remains that humans are headed to obsoleteness…again. Humans have faced this issue since the start of human ingenuity. The key is to adapt and realize what AI truly is. Elizabeth Mathew of Signoz.io wrote: “I Built An MCP Server For Observability. This Is My Unhyped Take.”
If you’re unfamiliar with an MCP server it is an open standard that defines how LLMS or AI agents (i.e. Claude) uniformly connect external tools and data sources. It can be decoupled and used similar to a USB-C device then used for any agent.
After explaining some issues with MCP servers and why they are “schizophrenic”,
Mathew concludes with this:
“Ultimately, MCP-powered agents are not bringing us closer to automated problem-solving. They are giving us sophisticated hypothesis generators. They excel at exploring the known, but the unknown remains the domain of the human engineer. We’re not building an automated SRE; we’re building a co-pilot that can brainstorm, but can’t yet reason. And recognizing that distinction is the key to using these tools effectively without falling for the hype.”
She might be true from an optimistic and expert perspective, but that doesn’t prevent CEOs from implementing AI to replace their workforce or young adults being encouraged away from coding careers. Recent college graduates, do you have a job, any job?
Whitney Grace, September 19, 2025
AI Search Is Great. Believe It. Now!
September 18, 2025
Sadly I am a dinobaby and too old and stupid to use smart software to create really wonderful short blog posts.
Cheerleaders are necessary. The idea is that energetic people lead other people to chant: Stand Up, Sit Down, Fight! Fight! Fight! If you get with the program, you stand up. You sit down. You shout, of course, fight, fight, fight. Does it help? I don’t know because I don’t cheer at sports events. I say, “And again” or some other statement designed to avoid getting dirty looks or caught up in standing, sitting, and chanting.
Others are different. “GPT-5 Thinking in ChatGPT (aka Research Goblin) Is Shockingly Good at Search” states:
Don’t use chatbots as search engines” was great advice for several years… until it wasn’t. I wrote about how good OpenAI’s o3 was at using its Bing-backed search tool back in April. GPT-5 feels even better.
The idea is that instead of working with a skilled special librarian and participating in a reference interview, people started using online Web indexes. Now we have moved from entering a query to asking a smart software system for an answer.
Consider the trajectory. A person seeking information works with a professional with knowledge of commercial databases, traditional (book) reference tools, and specific ways of tracking down and locating information needed to answer the user’s question. When the user was not sure, the special librarian would ask, “What specific information do you need?” Some users would reply, “Get me everything about subject X?” The special librarian would ask other questions until a particular item could be identified. In the good old days, special librarians would seek the information and provide selected items to the person with the question. Ellen Shedlarz at Booz, Allen & Hamilton when I was a lowly peon did this type of work as did Dominque Doré at Halliburton NUS (a nuclear outfit).
We then moved to the era of PCs and do-it-yourself research. Everyone became an expert. Google just worked. Then mobile phones arrived so research on the go was a thing. But keying words into a search box and fiddling with links was a drag. Now just tell the smart software your problem. The solution is just there like instant oatmeal.
The Stone Age process was knowledge work. Most people seeking information did not ask, preferring as one study found to look through trade publications in an old-fashioned in box or pick up the telephone and ask a person whom one assumed knew something about a particular subject. The process was slow, inefficient, and fraught with delays. Let’s be efficient. Let’s let software do everything.
Flash forward to the era of smart software or seemingly smart software. The write up reports:
I’ve been trying out hints like “go deep” which seem to trigger a more thorough research job. I enjoy throwing those at shallow and unimportant questions like the UK Starbucks cake pops one just to see what happens! You can throw questions at it which have a single, unambiguous answer—but I think questions which are broader and don’t have a “correct” answer can be a lot more fun. The UK supermarket rankings above are a great example of that. Since I love a questionable analogy for LLMs Research Goblin is… well, it’s a goblin. It’s very industrious, not quite human and not entirely trustworthy. You have to be able to outwit it if you want to keep it gainfully employed.
The reference / special librarians are an endangered species. The people seeking information use smart software. Instead of a back-and-forth and human-intermediated interaction between a trained professional and a person with a question, we get “trying out” and “accepting the output.”
I think there are three issues inherent in this cheerleading:
- Knowledge work is short circuited. Instead of information-centric discussion, users accept the output. What if the output is incorrect, biased, incomplete, or made up? Cheerleaders shout more enthusiastically until a really big problem occurs.
- The conditioning process of accepting outputs makes even intelligent people susceptible to mental short cuts. These are good, but accuracy, nuance, and a sense of understanding the information may be pushed to the side of the information highway. Sometimes those backroads deliver unexpected and valuable insights. Forget that. Grab a burger and go.
- The purpose of knowledge work is to make certain that an idea, diagnosis, research study can be trusted. The mechanisms of large language models are probabilistic. Think close enough for horseshoes. Cheering loudly does not deliver accuracy of output, just volume.
Net net: Inside each large language model lurks a system capable of suggesting glue cheese on pizza, the gray mass is cancer, and eat rocks.
What’s been lost? Knowledge value from the process of obtaining information the Stone Age way. Let’s work in caves with fire provided by burning books. Sounds like a plan, Sam AI-Man. Use GPT5, use GPT5, use GPT5.
Stephen E Arnold, September 18, 2025
AI Maggots: Are These Creatures Killing the Web?
September 18, 2025
The short answer is, “Yep.”
The early days of the free, open Web held such promise. Alas, AI is changing the Internet and there is, apparently, nothing we can do about it. The Register laments, “AI Web Crawlers Are Destroying Websites in their Never-Ending Hunger for Any and All Content: But the Cure May Ruin The Web.…” Writer Steven J. Vaughan-Nichols tells us a whopping 30% of traffic is now bots, according to Cloudflare. And 80% of that, reports Fastly, comes from AI-data fetcher bots. Web crawlers have been around since 1993, of course, but this volume is something new. And destructive. Vaughan-Nichols writes:
“Fastly warns that [today’s AI crawlers are] causing ‘performance degradation, service disruption, and increased operational costs.’ Why? Because they’re hammering websites with traffic spikes that can reach up to ten or even twenty times normal levels within minutes. Moreover, AI crawlers are much more aggressive than standard crawlers. As the InMotionhosting web hosting company notes, they also tend to disregard crawl delays or bandwidth-saving guidelines and extract full page text, and sometimes attempt to follow dynamic links or scripts. he result? If you’re using a shared server for your website, as many small businesses do, even if your site isn’t being shaken down for content, other sites on the same hardware with the same Internet pipe may be getting hit. This means your site’s performance drops through the floor even if an AI crawler isn’t raiding your website. Smaller sites, like my own Practical Tech, get slammed to the point where they’re simply knocked out of service. Thanks to Cloudflare Distributed Denial of Service (DDoS) protection, my microsite can shrug off DDoS attacks. AI bot attacks – and let’s face it, they are attacks – not so much.”
Even big websites are shelling out for more processor, memory, and network resources to counter the slowdown. And no wonder: According to Web hosting firms, most visitors abandon a site that takes more than three seconds to load. Site owners have some tools to try mounting a defense, like paywalls, logins, and annoying CAPTCHA games. Unfortunately, AI is good at getting around all of those. As for the tried and true, honor-system based robots.txt files, most AI crawlers breeze right on by. Hey, love maggots.
Cynthia Murrell, September 18, 2025
YouTube: Behind the Scenes Cleverness?
September 17, 2025
No smart software involved. Just a dinobaby’s work.
I read “YouTube Is a Mysterious Monopoly.” The author tackles the subject of YouTube and how it seems to be making life interesting for some “creators.” In many countries, YouTube is television. I discovered this by accident in Bucharest, Cape Town, and Santiago, to name three locations where locals told me, “I watch YouTube.”
The write up offers some comments about this Google service. Let’s look at a couple of these.
First, the write up says:
…while views are down, likes and revenue have been mostly steady. He guesses that this might be caused by a change in how views are calculated, but it’s just a guess. YouTube hasn’t mentioned anything about a change, and the drop in views has been going on for about a month.
About five years ago, one of the companies with which I have worked for a while, pointed out that their Web site traffic was drifting down. As we monitored traffic and ad revenues, we noticed initial stability and then a continuing decline in both traffic and ad revenue. I recall we checked some data about competitive sites and most were experiencing the same drift downwards. Several were steady or growing. My client told me that Google was not able to provide substantive information. Is this type of decline an accident or is it what I call traffic shaping for Google’s revenue? No one has provided information to make this decline clear. Today (September 10, 2025) the explanation is related to smart software. I have my doubts. I think it is Google cleverness.
Second, the write up states:
I pay for YouTube Premium. For my money, it’s the best bang-for-the-buck subscription service on the market. I also think that YouTube is a monopoly. There are some alternatives — I also pay for Nebula, for example — but they’re tiny in comparison. YouTube is effectively the place to watch video on the internet.
In the US, Google has been tagged with the term “monopoly.” I find it interesting that YouTube is allegedly wearing a T shirt that says, “The only game in town.” I think that YouTube has become today’s version of the Google online search service. We have people dependent on the service for money, and we have some signals that Google is putting its thumb on the revenue scale or is suffering from what users are able to view on the service. Also, we have similar opaqueness about who or what is fiddling the dials. If a video or a Web site does not appear in a search result, that site may as well not exist for some people. The write up comes out and uses the “monopoly” word for YouTube.
Finally, the essay offers this statement:
Creators are forced to share notes and read tea leaves as weird things happen to their traffic. I can only guess how demoralizing that must feel.
For me, this comment illustrates that the experience of my client’s declining traffic and ad revenue seems to be taking place in the YouTube “datasphere.” What is a person dependent on YouTube revenue supposed to do when views drop or the vaunted YouTube search service does not display a hit for a video directly relevant to a user’s search. OSINT experts have compiled information about “Google dorks.” These are hit-and-miss methods to dig a relevant item from the Google index. But finding a video is a bit tricky, and there are fewer Google dorks to unlock YouTube content than for other types of information in the Google index.
What do I make of this? Several preliminary observations are warranted. First, Google is hugely successful, but the costs of running the operation and the quite difficult task of controlling the costs of ping, pipes, and power, the cost of people, and the expense of dealing with pesky government regulators. The “steering” of traffic and revenue to creators is possibly a way to hit financial targets.
Second, I think Google’s size and its incentive programs allow certain “deciders” to make changes that have local and global implications. Another Googler has to figure out what changed, and that may be too much work. The result is that Googlers don’t have a clue what’s going on.
Third, Google appears to be focused on creating walled gardens for what it views as “Web content” and for creator-generated content. What happens when a creator quits YouTube? I have heard that Google’s nifty AI may be able to extract the magnetic points of the disappeared created and let its AI crank out a satisfactory simulacrum. Hey, what are those YouTube viewers in Santiago going to watch on their Android mobile device?
My answer to this rhetorical question is the creator and Google “features” that generate the most traffic. What are these programs? A list of the alleged top 10 hits on YouTube is available at https://mashable.com/article/most-subscribed-youtube-channels. I want to point out that the Google holds down position in its own list spots number four and number 10. The four spot is Google Movies, a blend of free with ads, rent the video, “buy” the video which sort of puzzles me, and subscribe to a stream. The number 10 spot is Google’s own music “channel”. I think that works out to YouTube’s hosting of 10 big draw streams and services. Of those 10, the Google is 20 percent of the action. What percentage will be “Google” properties in a year?
Net net: Monitoring YouTube policy, technical, and creator data may help convert these observations into concrete factoids. On the other hand, you are one click away from what exactly? Answer: Daily Motion or RuTube? Mysterious, right?
Stephen E Arnold, September 17, 2025
Desperate Much? Buying Cyber Security Software Regularly
September 16, 2025
Bad actors have access to AI, and it is enabling them to increase both speed and volume at an alarming rate. Are cybersecurity teams able to cope? Maybe—if they can implement the latest software quickly enough. VentureBeat reports, “Software Commands 40% of Cybersecurity Budgets ad Gen AI Attacks Execute in Milliseconds.” Citing IBM’s recent Cost of a Data Breach Report, writer Louis Columbus reports 40% of cybersecurity spending now goes to software. Compare that to just 15.8% spent on hardware, 15% on outsourcing, and 29% on personnel. Even so, AI-assisted hacks now attack in milliseconds while the Mean Time to Identify (MTTI) is 181 days. That is quite the disparity. Columbus observes:
“Three converging threats are flipping cybersecurity on its head: what once protected organizations is now working against them. Generative AI (gen AI) is enabling attackers to craft 10,000 personalized phishing emails per minute using scraped LinkedIn profiles and corporate communications. NIST’s 2030 quantum deadline threatens retroactive decryption of $425 billion in currently protected data. Deepfake fraud that surged 3,000% in 2024 now bypasses biometric authentication in 97% of attempts, forcing security leaders to reimagine defensive architectures fundamentally.”
Understandable. But all this scrambling for solutions may now be part of the problem. Some teams, we are told, manage 75 or more security tools. No wonder they capture so much of the budget. Simplification, however, is proving elusive. We learn:
“Security Service Edge (SSE) platforms that promised streamlined convergence now add to the complexity they intended to solve. Meanwhile, standalone risk-rating products flood security operations centers with alerts that lack actionable context, leading analysts to spend 67% of their time on false positives, according to IDC’s Security Operations Study. The operational math doesn’t work. Analysts require 90 seconds to evaluate each alert, but they receive 11,000 alerts daily. Each additional security tool deployed reduces visibility by 12% and increases attacker dwell time by 23 days, as reported in Mandiant’s 2024 M-Trends Report. Complexity itself has become the enterprise’s greatest cybersecurity vulnerability.”
See the writeup for more on efforts to improve cybersecurity’s speed and accuracy and the factors that thwart them. Do we have a crisis yet? Of course not. Marketing tells us cyber security just works. Sort of.
Cynthia Murrell, September 16, 2025
Shame, Stress, and Longer Hours: AI’s Gifts to the Corporate Worker
September 15, 2025
Office workers from the executive suites to entry-level positions have a new reason to feel bad about themselves. Fortune reports, “ ‘AI Shame’ Is Running Rampant in the Corporate Sector—and C-Suite Leaders Are Most Worried About Getting Caught, Survey Says.” Writer Nick Lichtenberg cites a survey of over 1,000 workers by SAP subsidiary WalkMe. We learn almost half (48.8%) of the respondents said they hide their use of AI at work to avoid judgement. The number was higher at 53.4% for those at the top—even though they use AI most often. But what about the generation that has entered the job force amid AI hype? We learn:
“Gen Z approaches AI with both enthusiasm and anxiousness. A striking 62.6% have completed work using AI but pretended it was all their own effort—the highest rate among any generation. More than half (55.4%) have feigned understanding of AI in meetings. … But only 6.8% report receiving extensive, time-consuming AI training, and 13.5% received none at all. This is the lowest of any age group.”
In fact, the study found, only 3.7% of entry-level workers received substantial AI training, compared to 17.1% of C-suite executives. The write-up continues:
“Despite this, an overwhelming 89.2% [of Gen Z workers] use AI at work—and just as many (89.2%) use tools that weren’t provided or sanctioned by their employer. Only 7.5% reported receiving extensive training with AI tools.”
So younger employees use AI more but receive less training. And, apparently, are receiving little guidance on how and whether to use these tools in their work. What could go wrong?
From executives to fresh hires and those in between, the survey suggests everyone is feeling the impact of AI in the workplace. Lichtenberg writes:
“AI is changing work, and the survey suggests not always for the better. Most employees (80%) say AI has improved their productivity, but 59% confess to spending more time wrestling with AI tools than if they’d just done the work themselves. Gen Z again leads the struggle, with 65.3% saying AI slows them down (the highest amount of any group), and 68% feeling pressure to produce more work because of it.”
In addition, more than half the respondents said AI training initiatives amounted to a second, stressful job. But doesn’t all that hard work pay off? Um, no. At least, not according to this report from MIT that found 95% of AI pilot programs at large companies fail. So why are we doing this again? Ask the investor class.
Cynthia Murrell, September 15, 2025
Common Sense Returns for Coinbase Global
September 5, 2025
No AI. Just a dinobaby working the old-fashioned way.
Just a quick dino tail slap for Coinbase. I read “Coinbase Reverses Remote Policy over North Korean Hacker Threats.” The write up says:
Coinbase has reversed its remote-first policy due to North Korean hackers exploiting fake remote job applications for infiltration. The company now mandates in-person orientations and U.S. citizenship for sensitive roles. This shift highlights the crypto industry’s need to balance flexible work with robust cybersecurity.
I strongly disagree with the cyber security angle. I think it is a return (hopefully) to common sense, not the mindless pursuit of cheap technical work and lousy management methods. Sure, cyber security is at risk when an organization hires people to do work from a far off land. The easy access to voice and image synthesis tools means that some outfits are hiring people who aren’t the people the really busy, super professional human resources person thinks was hired.
The write up points out:
North Korean hackers have stolen an estimated $1.6 billion from cryptocurrency platforms in 2025 alone, as detailed in a recent analysis by Ainvest. Their methods have evolved from direct cyberattacks to more insidious social engineering, including fake job applications enhanced by deepfakes and AI-generated profiles. Coinbase’s CEO, Brian Armstrong, highlighted these concerns during an appearance on the Cheeky Pint podcast, as covered by The Verge, emphasizing how remote-first policies inadvertently create vulnerabilities.
Close but the North Korean angle is akin to Microsoft saying, “1,000 Russian hackers did this.” Baloney. My view is that the organized hacking operations blend smoothly with the North Korean government’s desire for free cash and the large Chinese criminal organizations operating money laundering operations from that garden spot, the Golden Triangle.
Stealing crypto is one thing. Coordinating attacks on organizations to exfiltrate high value information is a second thing. A third thing is to perform actions that meet the needs and business methods of large-scale money laundering, phishing, and financial scamming operations.
Looking at these events from the point of view of a single company, it is easy to see that cost reduction and low cost technical expertise motivated some managers, maybe those at Coinbase. But now that more information is penetrating the MBA fog that envelopes many organizations, common sense may become more popular. Management gurus and blue chip consulting firms are not proponents of common sense in my experience. Coinbase may have seen the light.
Stephen E Arnold, September 5, 2025
Grousing Employees Can Be Fun. Credible? You Decide
September 4, 2025
No AI. Just a dinobaby working the old-fashioned way.
I read “Former Employee Accuses Meta of Inflating Ad Metrics and Sidestepping Rules.” Now former employees saying things that cast aspersions on a former employer are best processed with care. I did that, and I want to share the snippets snagging my attention. I try not to think about Meta. I am finishing my monograph about Telegram, and I have to stick to my lane. But I found this write up a hoot.
The first passage I circled says:
Questions are mounting about the reliability of Meta’s advertising metrics and data practices after new claims surfaced at a London employment tribunal this week. A former Meta product manager alleged that the social media giant inflated key metrics and sidestepped strict privacy controls set by Apple, raising concerns among advertisers and regulators about transparency in the industry.
Imagine. Meta coming up at a tribunal. Does that remind anyone of the Cambridge Analytica excitement? Do you recall the rumors that fiddling with Facebook pushed Brexit over the finish line? Whatever happened to those oh-so-clever CA people?
I found this tribunal claim interesting:
… Meta bypassed Apple’s App Tracking Transparency (ATT) rules, which require user consent before tracking their activity across iPhone apps. After Apple introduced ATT in 2021, most users opted out of tracking, leading to a significant reduction in Meta’s ability to gather information for targeted advertising. Company investors were told this would trim revenues by about $10 billion in 2022.
I thought Apple had their system buttoned up. Who knew?
Did Meta have a response? Absolutely. The write up reports:
“We are actively defending these proceedings …” a Meta spokesperson told The Financial Times. “Allegations related to the integrity of our advertising practices are without merit and we have full confidence in our performance review processes.”
True or false? Well….
Stephen E Arnold, September 4, 2025
-
- Subscribe to Beyond Search
Feature archive
News archive
-
