The PJM Part 2 … Negative Externalities, Data Center Alley, and future historically high electricity bills throughout the PJM

Last update:  7:36am Wednesday 11/26/25

This note, Part 2 in what was originally planned to be a three part series, will now be the final installment. It highlights some operational structures of the PJM that facilitated the stunning success of Virginia’s Data Center Alley. Then it sketches a strategy that might enable the founding members of the PJM -- Pennsylvania, New Jersey, and Maryland -- to join Virginia as major national and international data center hubs via consortiums that would also guarantee stable lower electricity bills for their regular customers, i.e., households, businesses, schools, government agencies, etc. 

Unfortunately, recent surprising announcements from two powerful utilities — Dominion and Exelon — threaten to burden regular customers throughout the 13 states in the PJM with historically high electricity bills.  Exelon’s announcement requires extensive revisions to Section F and Dominion’s requires the addition of a forthcoming Section G to this note.

Op-Ed ➡ HERE

TL;DR 28Oct25 ... (1) How China Raced Ahead of the U.S. on Nuclear Power, (2) Anthropic's chatbot can now remember your previous conversations without prompting -- like ChatGPT and Gemini, and (3) OpenAI’s Atlas Browser Takes Direct Aim at Google

Last update: Tuesday 10/28/25

Welcome to our 28Oct25 TL;DR summaries by ChatGPT of the past week's top 3 stories on our "Useful AI News" page ➡  (1) How China Raced Ahead of the U.S. on Nuclear Power, (2) Anthropic's chatbot can now remember your previous conversations without prompting -- like ChatGPT and Gemini, and (3) OpenAI’s Atlas Browser Takes Direct Aim at Google Chrome

TL;DR ➡ HERE

What is the PJM Interconnect? How does it work? And why you should care? ... 1-Basics

Last update:  Saturday 3:27am 10/25/25

This note, Part 1 in a three part series, will provide an introduction to the PJM Interconnect that directs the transmission of electric power across thirteen states and the District of Columbia. Part 2 will highlight features of the PJM that provide persistent advantages — lower electricity bills — for households and enterprises in some states, while imposing persistent penalties — higher electricity bills — on households and enterprises in other states. Part 3 will analyze the ironic causes of the persistent penalties, i.e., the negative externalities,  that are being reported by the media with increasing frequency and stridency. 

Click HERE to view full report

A closer look at the potential world-wide benefits of Apple Silicon

Substantial Revisions: 7:19pm Friday 10/24/25

This second note in an occasional series explores an emerging solution to a global disaster. It takes a closer look at the potential world-wide benefits of Apple’s forthcoming sales of energy-efficient inference servers to its customers. It also corrects a few misconceptions in the first note 

To read more of this blog note click ➡ HERE

TL;DR 22Sep25 ... (1) China bans tech companies from buying Nvidia’s AI chips, (2) Nvidia to Invest $5 Billion in Intel, and (3) Harvard researchers weigh AI’s climate and health impact

Last update: Monday 9/22/25

Welcome to our 22Sep25 TL;DR summaries by ChatGPT of the past week's top 3 stories on our "Useful AI News" page ➡ (1) China bans tech companies from buying Nvidia’s AI chips, (2) Nvidia to Invest $5 Billion in Intel, and (3) Harvard researchers weigh AI’s climate and health impact

 

TL;DR ➡ HERE

TLDR 15Sep25 ... (1) Apple Introduces New, Slimmer iPhone, (2) Microsoft to Buy AI From Anthropic, and (3) EPA Proposes to End Greenhouse Gas Reporting Program

Last update: Monday 9/15/25

Welcome to our 15Sep25 TL;DR summaries by ChatGPT of the past week's top 3 stories on our "Useful AI News" page ➡  (1) Apple Introduces New, Slimmer iPhone, (2) Microsoft to Buy AI From Anthropic, and (3) EPA Proposes to End Greenhouse Gas Reporting Program

 

TL;DR ➡ HERE

TL;DR 1Sep25 ... (1) AI super PACs, (2) Nvidia Predicts Cooler Growth After Sales Record, and (3) Data centers will cause higher electricity prices

Last update: Monday 9/1/25

Welcome to our 1Sep25 TL;DR summaries by ChatGPT of the past week's top 3 stories on our "Useful AI News" page ➡  (1) AI super PACs, (2) Nvidia Predicts Cooler Growth After Sales Record, and (3) Data centers will cause higher electricity prices

 

TL;DR ➡  HERE

Sometimes ChatGPT does not connect the dots. Sometimes it does not follow its "plans".

Last update: Tuesday 8/26/25

Everybody knows that chatbotmake mistakes. This misbehavior has given rise to the fundamental guideline that users should always doublecheck everything a chatbot says. The  implicit assumption of this guideline is that a chatbot’s incorrect response to a prompt, a/k/a an “error” reflects the chatbot’s incorrect knowledge. This note challenges that assumption. 

The editor of this blog can document situations wherein ChatGPT “knew” the correct response to a prompt, but delivered an incorrect response anyway. In other words, ChatGPT failed to connect the dots, failed to give the response that it “knew” that it should have given. 

To read more of this blog note click ➡ HERE

TL;DR ... 18Aug 25 ... (1) Question: Does Nvidia sell chips to China: Yes or No? ... Trump says "Yes", China says "No" ... and (2) Local negative consequence of the data center boom

Last update: Tuesday 8/19/25

Welcome to our 18Aug25 TL;DR summaries by ChatGPT of the past week's top 2 stories on our "Useful AI News" page ➡  (1) Question: Does Nvidia sell chips to China: Yes or No? ... Trump says "Yes", China says "No" ... and (2) Local negative consequence of the data center boom: Higher electricity bills for everyone and lower air quality

TL;DR ➡ HERE

Apple moves back to its roots in hardware with “cloud buster” chips

Extensive revision: 8/17/25

This is the first of a series of occasional publications that will track Apple’s development of genAI inference data centers using Apple Silicon, centers at the edges of the Internet that will support billions of devices. Given Apple's policy of minimal publicity during development, there probably won't be more than four more reports before the end of 2026.

Note: During “training”, a model learns from a large dataset; during “inference”, a trained model responds to queries from users.

To read more of this blog note click ➡  HERE

The Great AI Exodus: How Three Tech Giants Lost the Visionaries Who Built the Future

Last update: Friday 8/1/25

An exclusive article in The Information last week reported that many members of Apple's large language model unit were leaving because they perceived that Apple's leadership did not share their vision of the transcendent importance of generative AI. Given that The Information's report was behind a very high paywall, most of the readers of this blog probably could not read it. So the editor tasked ChatGPT to provide an extensive TLDR summary  

To read more of this blog note click ➡ HERE

Gemini ... The world's first genAI browsing tool

Last update: Saturday 7/26/25

In a recent note on this blog we ranked the three leading chatbots -- ChatGPT, Claude, and Gemini -- with regards to 8 basic functions. None of the three was the best at everything; and Gemini only excelled at two things: searching the Internet and following links to unrestricted pages (e.g., pages not behind paywalls). However, its dominance of these two functions enables it to become the world's first genAI browsing tool, a new function that will be discussed in this note ... Click HERE to read more.

Once upon a recent time, there was a race between a tortoise named ChatGPT, a tortoise named Claude, and a very fast hare named Gemini

Last update: Wednesday 7/16/25

Once upon a recent time, there was a race between a tortoise named ChatGPT, another tortoise named Claude, and a fast moving hare named Gemini. At the start of the race, everybody knew who was supposed to win. But as quick as a wink, some unexpected things began to happen. 
 

Click HERE to read more. 

Note: The cartoon image for this note was composed via ChatGPT on GPT-4o.

The Anthropic tortoise and the Google hare

Last update: Friday 7/11/25

Once upon a recent time, the editor of this blog posed two questions to Anthropic’s Claude and to Google’s Gemini. Their responses were remarkably similar.

Q1. Why is Google playing catch up to Anthropic in generative AI, a game that Google invented? 

Q2. How come Anthropic, a small startup, has done more significant research in the last three years about how generative AI models really work than Google, whose generative AI efforts are led by Dr. Demis Hassabis, a researcher whose indisputable brilliance earned him a well deserved Nobel prize in chemistry in 2024 for his previous AI innovations.
Click Here to read more

Note: The cartoon image for this note was composed via ChatGPT on GPT-4o.

Powerful AI data centers in your home/office using Apple technology, Part 2 -- How to build a data center with no code, with no code, with no code

Friday 7/4/25

Your interactions with chatbots, agents, and other AI apps reveal how you really think, especially as these apps become more conversational. Knowing how you really think will enable advertisers to design sales pitches that emphasize features that you are most likely to find appealing in their efforts to persuade you to buy their products and services. These revelations can also enable intrusive government agencies and insurance companies to become even more intrusive. 

Big Tech companies compile insights about you -- which they derive from your interactions with their cloud-based, large language models -- into encyclopedic profiles. They sell access to these profiles to the highest bidders. An obvious and powerful way to substantially reduce these intrusions is to redirect a substantial portion of your queries away from Big Tech's clouds ... Click HERE to read more

 

Powerful AI data centers in the home/office using Apple technology, Part 1 -- Greatly reduced carbon footprints

Last update: Saturday 6/28/25

Having a personal generative AI data center in your home or office confers two obvious benefits ➡ (1) substantial reduction of your carbon footprint, and (2) greater privacy. This note estimates the reduction in the size of your carbon footprint that can be achieved by using Apple technology to converse with powerful, open source, small language models in your local center vs. large language models in the cloud. A second note will describe how the editor of this blog built his own AI datacenter using Apple technology ... with no code.
Click HERE to read more

GenAI Diary (page) ...The absence of critical thinking in the research design of a recent MIT Media Lab study of student misuse of ChatGPT

Last update: Sunday 6/22/25

This note discusses the absence of critical thinking in the research design of a recent MIT Media Lab study of student misuse of ChatGPT.  

Apple's dilemma: Pleasing its iPhone users AND its long term investors

Last update: Tuesday 7/8/25

   Apple Park, Cupertino, CA

Apple Park — corporate headquarters for Apple, Inc — is not in the real world. Its unreality emerged in WWDC 2024 and has just been confirmed by WWDC 2025. But as this note will argue, it doesn't matter. Apple merely needs to follow the pragmatic build AND buy strategy that it has followed with great success since 1979. Time is not on Apple's side; but its $3.5 trillion dollar market cap will win the day. It just has to make one more very large, high tech acquisition within the next two years. But which one? ... Click  HERE  to read more.

How and why we should build a national library of generative AI use cases and benchmark tests

Last update: Monday 6/9/25

Coding has rapidly become the most successful use case for Large Language Models. Models can now code so well that all Big Tech and many smaller companies are making drastic reductions in the size of their coding staffs. That's bad news ... for software engineers. 

The good news ... for experts in a wide range of other fields ... is that new employment opportunities are just over the horizon for computer savvy personnel who learn how to use generative AI to solve complex problems in their fields. So how can they learn these new skills? Click Here to read more ... 

GenAI Diary (page) ... Major AI Research Breakthroughs and Reasoning Models (2017–Early 2025)

Last update: Tuesday 4/29/25

Last week, Dario Amode, Anthropic's CEO, published a plea to his colleagues in the genAI community to join Anthropic’s quest to understand the inner workings of large language models. Amodei warned that genAI developers were building smarter models at a much faster pace than the pace at which their understanding was increasing of how these black box models really worked. Although Anthropic had recently mapped millions of neurons to specific concepts, this achievement must be measured against the billions of other neural clusters in their models for which Anthropic still has no understanding  

Podcast ... Using generative AI agents to make Signal a safe and secure administrative tool for all presidents ... Part 1, SignalGate

Last update: Wednesday 4/9/25 

Welcome to the 1st of a series of 4 podcasts about the Signal app. The series is called, "Using generative AI agents to make Signal a safe and secure administrative tool for all presidents". This first podcast answers some questions raised by our readers, friends, and associates about the issues addressed in the lengthy report that we posted last week, "SignalGate ... Alternative timelines of a leaky chat". This fiasco demonstrated how badly things can go when high level officials with insufficient training and insufficient technical support use a world class open source messaging app to discuss highly confidential subjects. ... The first section of this podcast is based on ChatGPT's succinct responses to a series of questions of concern to potential new users of Signal. What’s it all about? The second section is based on the chatbot's TL;DR summary of our lengthy report about the fiasco's double-hack. As usual, our podcast audio is a recording of the Descript app reading the chatbot's responses and summaries in an AI imitation of the voices of the editor of this blog and his AI colleagues.

Display transcript ➡ HERE

Audio podcast ...17min 

 

If audio fails to start, or gets stuck, try reloading the page

SignalGate ... Alternative timelines of a leaky chat

Last update: Sunday 4/6/25

Here's the headline: House Democratic leader "Jeffries calls for firing Hegseth over war plans group chat", Nicholas Wu, Politico, 3/25/25 ... This blog note offers an alternative interpretation of what happened, an alternative that yields two recommendations that might minimize the chances of this happening again by putting GenAI agents on center stage.