Ken's Korner Newsletter Logo July 2026
Where Aare We Going We Going With This AI Thing

Part Two

Welcome to the September - October 2026 issue of Ken’s Korner Newsletter

Original Processor.

Earlier this year Ken’s Korner published a newsletter asking questions about artificial intelligence. This month we are going to revisit that subject with …
Where are we going with this AI thing now?

The short answer is, “We are going broke!”. “We just hope that the other guys go broke before we do.”.

AI makes a lot of mistakes.
There has been much talk lately about how artificial intelligence is going to take over the world. Humans will be obsolete and our civilization is doomed. A great deal of “gloom and doom” talk about all the bad things that AI could do has suddenly sprung up. You can’t watch a newscast and not catch some of that.

Remember that this is an election year and a number of governments, including our own, have members that think the government can cure all ills. Just give them full control over everything.

In my experience AI has often been a remarkable failure. Recently I had to replace the sound bar on an entertainment center. Perhaps not the normal sort of job that people imagine when they think about an Information Technology tech, but what the heck. Work is work, right.

It should have been a cakewalk, but it turned out to be a several days long process. The old 5.1 sound system was nearly twenty years old and no longer worked. It was state-of-the-art back then, but the times have changed and the new systems are a vast improvement.

The instructions were very simple, plug the new sound bar in and turn it on. One power cable and one HDMI cable. Problem, no sound comes out of either the sound bar or the bass unit, (woofer). The first thing to do is check all the connections. With one HDMI port and one power plug and yes, the little power light is on, that part was easy.

The sound bar is connected to the HDMI port on the one power plug and HDMI port labeled ARC, (Audio Return Chanel) as specified in the instructions. But this System isn’t running as a conventional TV. It is more like a big monitor for the PC. Still, it is connected via HDMI and by the HDMI spec the audio should be included

So, I did a Google search, which means that the AI, in this case was Gemini, I gave it the make and model number of the units in use and a description of the problem, no sound. It had me go through a repeat of the steps that I had already taken. Not much help there and still no sound. I told Gemini that there was still no sound and it suggested that it might have a problem with the handshake. Possible, but HDMI is a solid, well tested and refined technology. I have never seen a handshake problem with HDMI that a restart or, at most a power cycle, didn’t fix.

Not wanting to leave any stone unturned I followed the AI’s instructions. The first thing was to change the power-on sequence. I turned on the sound bar and waited until it had fully booted up. Then I turned on the TV and waited for it to fully boot up and last I turned on the PC that provides input to the TV. Still no sound. I saw some article that had that order reversed so the sound bar was the last thing to get turned on. I tried both ways. Still no sound.

At this point I was ready to give up on the idea of an HDMI connection. Too many pieces didn’t fit together. Yes, the audio is supposed to be carried on the HDMI. But this isn’t a typical TV setup. The TV is being used as a big monitor, not as a TV. Although it is supplied by an HDMI connection and it does have an ARC type HDMI port the HDMI is from a PC. In theory this should work fine.

My other option was to connect via Bluetooth. Beng a glutton for punishment I went back to the AI for a step-by-step process for connecting the sound bar with Bluetooth. Again, I gave it the make and model of the sound bar.

The AI, again Gemini, told me to set the PC in search mode and activate Bluetooth on the sound bar by pressing the paring button on the back of the sound bar. Problem, there is no button on the back of the sound bar. There are no external buttons or switches anywhere on the sound bar. Everything is done through the App and the remote control.

Oh, you are right replied the AI as it hastily put up another set of instructions. This model does not have external controls. At this point would you believe anything from this AI? I wasn’t! I just used my basic instinct and experience and connected the sound bar via Bluetooth. This PC didn’t have native Bluetooth. So, I added a Bluetooth dongle to one of the USB ports. We had great sound about fifteen minutes later. (That included downloading the required software and rebooting the computer).

Gemini has had problems since the beginning. Remember when it pictured Abe Lincoln as a black man and showed us Asian Nazis as part of history. It was so bad that Google issued an apology and took the image generation feature offline for some time while they fixed some of the underlying problems.

One of the main reasons that Gemini, and other AI models, makes so many mistakes is because it is a statistical prediction engine that guesses the next words based on pattern rather than really understanding the facts. Monkey see – monkey do.

But Gemini is not alone in the stupid AI storage bin. Recently I had some fun with Chat GPT. The task was to create an image or picture using only text-based instructions. OK so it was many pages of text over several days.

Chat GPT had a lot of trouble following instructions. I would tell it to change this item and not to make any other changes. It made the requested change and then changed two other items. When confronted with this behavior it simply said oh yea, sorry about that. It seems like you are dealing with a petulant child. Except how do you discipline a misbehaving AI?

IT Super Hero.

Here is the result. A toy superhero computer geek. Carded in a blister pack with a laptop and a little dog Ready to go on the shelf at your local toy store next to Spiderman and Superman. It looks real but it is all just a fantasy.

I could go on for some time just on the mistakes that I have personally seen AI make but I think it is time to branch off here and look at some different AI missteps.

AI and Fake News.
In September 2026 an analyst from U.S Special Operations Command used an AI chatbot to interpret open-source data and classified signals intelligence regarding a Chinese commercial vessel moving through the Middel East. The chatbot hallucinated and erroneously concluded that the ship was carrying nuclear material for Iran.

The false report circulated through the military like wildfire. Thinking that the threat was real, the US military prepared an interdiction operation. Armed boarding teams with air support were scrambled to the area.

Just minutes before the operation was to begin officials realized the report was flawed and the mission was called off. No one boarded the vessel.

The situation ended peacefully but the danger here is obvious. The hallucinating AI could have started a severe international crisis, even an armed conflict between the superpowers.

And there is more, much more.
You have probably heard that vast sums of money are being spent on Artificial Intelligence. There is no shortage of news reports about how much it costs to set up an AI model, train the model, create a datacenter to house that model or huge amount of electrical power needed to build and run AI. But there are other ways that AI can cost a lot of money.

Google Bard Alphabet Inc. (The parent company of Google and Google Bard) lost $100 Billion in market cap on February 8, 2023, when its new chatbot Bard made a factual error during a promotional video demonstration. The question was asked, “What new discoveries from the James Webb Space Telescope can I tell my nine-year-old about?

Bard claimed that the telescope took the very first pictures of an exoplanet. In fact, the first pictures of an exoplanet came from the European Southern Observatory’s Very Larg Telescope, (VLT) way back in 2004.

Money for AI Zillow Offers $407.9 million lost after Zillow shut down its automated home-flipping business. The AI pricing algorithm misjudged housing market volatility and home valuations forcing the sale of thousands of homes at a steep loss

VW Cariad $7.5 Billion overrun. Volkswagen’s software division suffered massive delays and budget blowouts trying to build a unified in-house operating system and AI architecture across 12 automotive brands.

IBM Watson Health $4 billion write-down. IBM poured billions into building an AI driven healthcare and cancer treatment advisory tool. They never got it to work properly and ended up selling the unit after failing to prove clinical and commercial viability.

Arup Deepfake Heist $25.6 Million. The global engineering firm Arup was the victim of a very sophisticated scam. A finance employee in Hong Kong received an email ostensibly from Arup’s UK based Chief Financial Officer requesting a confidential and secret transaction and urging great discretion.

While initially skeptical the employee joined a live video conference call to discuss the matter. On the call the employee saw and heard what appeared to be the CFO and several trusted advisors from the London Office. In fact, every person on the call except the Hong Kong employee was an AI generated deepfake.

Convinced by the realistic video the employee executed fifteen separate wire transfers totaling HK$200 million, (about $25.6 million US) across five local Hong Kong bank accounts. The fraud was discovered several days later when the employee casually checked in with home office and learned that no such transaction has ever been authorized.

Also, this does not even begin to cover the environmental and resource toll that comes with AI. There has been no shortage of reports on costs of valuable commodities like electricity, water and real estate either.

If artificial Intelligence is going to take over the world and make humans obsolete it is going to need to do a much better job of getting it right on the first attempt. Otherwise, AI may find itself replaced by the next big thing to come along.

If you know someone that you think would enjoy this newsletter, share it with them and ask them to join using the link at the bottom of the page.

 

And remember — always back it up!

 

 

Go back to the top

To get the Ken's Korner Newsletter delivered to your Inbox CLICK HERE