Yep, the Old Internet Is Gone. Learn to Love the New Internet

August 1, 2024

green-dino_thumb_thumbThis essay is the work of a dumb humanoid. No smart software required.

The market has given the Google the green light to restrict information. The information highway has a new on ramp. If you want content created by people who were not compensated, you have to use Google search. Toss in the advertising system and that good old free market is going to deliver bumper revenue to stakeholders.

image

Online search is a problem. Here’s an old timer like me who broke his leg. The young wizard who works at a large online services firm explains that I should not worry. By the time my leg heals, I will be dead. Happy thoughts from one of those Gen somethings. Thanks, MSFT Copilot. How your security systems today?

What about users? The reality is that with Google the default search system in Apple iPhones, the brand that has redefined search and retrieval to mean “pay to play,” what’s the big deal?

Years ago I explained in numerous speeches and articles in publications like Online Magazine that online fosters the creation of centralized monopolistic information services. Some information professionals dismissed my observation as stupid. The general response was that online would generate benefits. I agree. But there were a few downsides. I usually pointed to the duopoly in online for fee legal information. I referenced the American Chemical Society’s online service Chemical Abstracts. I even pointed out that outfits like Predicasts and the New York Times would have a very, very tough time creating profitable information centric standalone businesses. The centralization or magnetic pull of certain online services would make generating profits very expensive.

So where are we now? I read “Reddit, Google, and the Real Cost of the AI Data Rush.” The article is representative of “real” journalists’, pundits’, and some regulators’ understanding of online information. The write up says:

Google, like Reddit, owes its existence and success to the principles and practices of the open web, but exclusive arrangements like these mark the end of that long and incredibly fruitful era. They’re also a sign of things to come. The web was already in rough shape, reduced over the last 15 years by the rise of walled-off platforms, battered by advertising consolidation, and polluted by a glut of content from the AI products that used it for training. The rise of AI scraping threatens to finish the job, collapsing a flawed but enormously successful, decades-long experiment in open networking and human communication to a set of antagonistic contracts between warring tech firms.

I want to point out that Google bought rights to Reddit. If you want to search Reddit, you use Google. Because Reddit is a high traffic site, users have to use Google. Guess what? Most online users do not care. Search means Google. Information access means Google. Finding a restaurant means Google. Period.

Google has become a center of gravity in the online universe. One can argue that Google is the Internet. In my monograph Google Version 2.0: The Calculating Predator that is exactly what some Googlers envisioned for the firm. Once a user accesses Google, Google controls the information world. One can argue that Meta and TikTok are going to prevent that. Some folks suggest that one of the AI start ups will neutralize Google’s centralized gravitational force. Google is a distributed outfit. Think of it as like the background radiation in our universe. It is just there. Live with it.

Google has converted content created by people who were not compensated into zeros and ones that will enhance its magnetic pull on users.

Several observations:

  1. Users were so enamored of a service which could show useful results from the quite large and very disorganized pools of digital information that it sucked the life out of its competitors.
  2. Once a funding source got the message through to the Backrub boys that they had to monetize, the company obtained inspiration from the Yahoo pay to play model which Yahoo acquired from Overture.com, formerly GoTo.com. That pay to play thing produces lots of money when there is traffic. Google was getting traffic.
  3. Regulators ignored Google’s slow but steady march to information dominance. In fact, some regulatory professionals with whom I spoke thought Google was the cat’s pajamas and asked me if I could get them Google T shirts for their kids. Google was not evil; it was fund; it was success.
  4. Almost the entire world’s intelligence professionals relay on Google for OSINT. If you don’t know what that means, forget the term. Knowing the control Google can exert by filtering information on a topic will probably give you a tummy ache.

The future is going to look exactly like the world of online in the year 1980. Google and maybe a couple of smaller also rans will control access to digital information. To get advertising free and to have a shot at bias free answers to online queries, users will have to pay. The currency will be watching advertising or subscribing to a premium service. The business model of Dialog Information Services, SDC, DataStar, and Dialcom is coming back. The prices will inflate. Control of information will be easy. And shaping or weaponizing content flow from these next generation online services will be too profitable to resist. Learn to love Google. It is more powerful than a single country’s government. If a country gets too frisky for Google’s liking, the company has ways to evade issues that make it uncomfortable.

The cartoon in this blog post summarizes my view of the situation. A fix will take a long time. I will be pushing up petunias before the problems of online search and the Information Superhighway are remediated.

Stephen E Arnold, August 1, 2024

Every Cloud Has a Silver Lining: Cyber Security Software from Israel

August 1, 2024

dinosaur30a_thumb_thumb_thumb_thumb__thumbThis essay is the work of a dinobaby. Unlike some folks, no smart software improved my native ineptness.

I wonder if those lucky Delta passengers have made it to their destinations yet? The Crowdstrike misstep caused a bit of a problem for some systems and for humans too. I saw a notice that CrowdStrike, founded by a Russian I believe, offered $10 to each person troubled by the teenie tiny mistake. Isn’t that too much for something which cannot be blamed on any one person, just on an elusive machine-centric process that had a bad hair day? Why pay anything?

And there is a silver lining to the CrowdStrike cloud! I read “CrowdStrike’s Troubles Open New Doors for Israeli Cyber Companies.” [Note that this source document may be paywalled. Just a heads up, gentle reader.] The write up asserts:

For the Israeli cyber sector, CrowdStrike’s troubles are an opportunity.

Yep, opportunity.

The write up adds:

Friday’s [July 26, 2024] drop in CrowdStrike shares reflects investor frustration and the expectation that potential customers will now turn to competitors, strengthening the position of Israeli companies. This situation may renew interest in smaller startups and local procurement in Israel, given how many institutions were affected by the CrowdStrike debacle.

The write up uses the term platformization, which is a marketing concept of the Palo Alto Networks cyber security firm. The idea is that a typical company is a rat’s nest of cyber security systems. No one is able to keep the features, functions, and flaws of several systems in mind. When something misfires or a tiny stumble occurs, Mr. Chaos, the friend of every cyber security professional, strolls in and asks, “Planning on a fun weekend, folks?”

image

The sales person makes reality look different. Thanks, Microsoft Copilot. Your marketing would never distort anything, right?

Platformization sounds great. I am not sure that any cyber security magic wand works. My econo-box automobile runs, but I would not say, “It works.” I can ponder this conundrum as I wait for the mobile repair fellow to arrive and riding in an Uber back to my office in rural Kentucky. The rides are evidence that “just works” is not exactly accurate. Your mileage may vary.

I want to point out that the write up is a bit of content marketing for Palo Alto Networks. Furthermore, I want to bring up a point which irritates some of my friends; namely, the Israeli cyber security systems, infrastructure, and smart software did not work in October 2023. Sure, there are lots of explanations. But which is more of a problem? CrowdStrike or the ineffectiveness of multiple systems?

Your call. The solution to cyber issues resides in informed professionals, adequate resources like money, and a commitment to security. Assumptions, marketing lingo, and fancy trade show booths simply prove that overpromising and under delivering is standard operating procedure at this time.

Stephen E Arnold, August 1, 2024

A Reliability Test for General-Purpose AI

August 1, 2024

A team of researchers has developed a valuable technique: “How to Assess a General-Purpose AI Model’s Reliability Before It’s Deployed.” The ScienceDaily article begins by defining foundation models—the huge, generalized deep-learning models that underpin generative AI like ChatGPT and DALL-E. We are reminded these tools often make mistakes, and that sometimes these mistakes can have serious consequences. (Think self-driving cars.) We learn:

“To help prevent such mistakes, researchers from MIT and the MIT-IBM Watson AI Lab developed a technique to estimate the reliability of foundation models before they are deployed to a specific task. They do this by considering a set of foundation models that are slightly different from one another. Then they use their algorithm to assess the consistency of the representations each model learns about the same test data point. If the representations are consistent, it means the model is reliable. When they compared their technique to state-of-the-art baseline methods, it was better at capturing the reliability of foundation models on a variety of downstream classification tasks. Someone could use this technique to decide if a model should be applied in a certain setting, without the need to test it on a real-world dataset. This could be especially useful when datasets may not be accessible due to privacy concerns, like in health care settings. In addition, the technique could be used to rank models based on reliability scores, enabling a user to select the best one for their task.”

Great! See the write-up for the technical details behind the technique. This breakthrough can help companies avoid mistakes before they launch their products. That is, if they elect to use it. Will organizations looking to use AI for cost cutting go through these processes? Sadly, we suspect that, if costs go down and lawsuits are few and far between, the AI is deemed good enough. But thanks for the suggestion, MIT.

Cynthia Murrell, August 1, 2024

  • Archives

  • Recent Posts

  • Meta