Category Archives: Technology

The 1-Step Guide to Responsible AI in Procurement

Forbes recently published an article on Responsible AI Procurement: A Practical Guide For Selecting Trustworthy AI Vendors. It wasn’t bad, but it missed the point.

Today, there’s only one way to responsibly address AI in Procurement.

JUST SAY NO!

1) We don’t really understand proper AI Governance (especially when most vendors are using third parties which are illegally scarping content, not checking for bias, and tweaking models on the fly without consideration for the new problems the on-the-fly tweaks will cause).

Plus, it’s not just ethical codes of conduct, it’s agreeing on what the ethics are, and, most importantly, making sure the models are transparent and unbiased — but we don’t know how to do that today, especially since all these models are huge black box models.

2) You can demand all the evidence you want from the vendor as backup for the vendor claims, but if you can’t verify it, how can you trust it?

3) These models require huge datasets to train. Even if you know the data set used and the processing method used, how can you be sure every element was properly vetted? Just like one bad apple can spoil the bunch, just one bad element in a clustering or optimization model can spoil the entire model. Just one!  It only takes a small amount of bad data to spoil a model, regardless of the model used.

4) These models can fail, and sometimes fail spectacularly. If you don’t understand the model, you don’t understand where it can fail, and thus what to look for. Also, many minor incidents (which can foretell future catastrophic failures) will go unnoticed if a human isn’t checking everything.

5) These models are not secure … the AI can leak any training data at any time without warning. Your vendor can have every security certification under the sun, and all will be for naught if they use LLMs.

So, JUST SAY NO!

A Circular Battery Economy is Necessary For Green Vehicles

Regulation (EU) 2023/1542 of the European Parliament and of the Council concerning batteries and waste batteries took effect this summer (on 12 July 2023 to be precise) and it’s a good first step towards a sustainable battery economy that will, hopefully, reduce carbon in the long-term.

When you consider that all of the zero emission claims for battery-powered vehicles are complete bullcr@p when you consider the carbon emissions to produce the vehicle, the carbon emissions required to produce the battery (which, in an inefficient process, can be more than 2X the emissions to produce the rest of the vehicle), and then the emissions to charge the battery from what is usually an oil or coal power plant, you might have to drive as much as 1,000,000 kms just to reach carbon neutrality! (And while the linked article doesn’t work out the best case scenario, it’s likely you’re driving the full warranty, or about 200,000 kms, to reach carbon neutrality when you’re charging the batteries burning oil or dirty coal.)

And even if the vehicle production is optimized, the battery production is optimized, and the power grid is primarily powered by pure renewable energy, it’s still not zero emission. The production of solar panels emit carbon, the production of windmills produce carbon, the building of dams and the generators that run them produce carbon, so you have to amortize that over the expected lifetime every time you charge that battery. So even a green vehicle will produce thousands of kilograms of carbon in its production, thousands of kilograms of carbon in its battery production, and hundreds to thousands of kilograms of carbon during its recharging. If you’re lucky enough to have the best case scenario with access to high efficiency solar, then you can get your carbon footprint down to about 10g per kwH over its expected lifetime, or a mere kg of carbon per full charge (or 400 kg over the first 200,000 km), and you approach carbon neutrality not long after you negate the production costs (which you might never do today as some methods to produce new batteries are so dirty). (But most of us do not have access to clean solar grids.)

This means that most first time produced vehicles with first time produced batteries are actually quite dirty. Very, very dirty. And the only way we’re ever going to get greener vehicles is to 1) cut down the carbon on vehicle production and 2) cut down the carbon in our power generation. Now, until we ban power production from oil, coal, and natural gas for fixed location power production (and build enough renewable power plants or start building micro modular reactor grids [where you could literally keep enough concrete on site to safely bury one in the case of a pending meltdown] and take advantage of the Onkalo spent nuclear fuel repository), there’s not much we can do about 2), but there are lots of things we can do about 1). First of all, we can make vehicles with more longevity (better part quality, more rustproof materials, easy part replacement, design for recycling, etc.). Secondly, we can design our batteries for reconditioning and recycling, to minimize the carbon production in the creation of future batteries, to make the next generation of vehicles greener.

But history has taught us no one does the up front research to design for recycling, or invests in recycling without regulation, so any regulation that forces companies to make more sustainable, circular, and safe batteries is not only a good thing, but the necessary first step on the road to truly getting green(er) vehicles.

Yes, McKinsey This Is Generative AI’s break out year, BUT:

We should NOT be celebrating the fact that it broke out of the prison it should be contained in only to:

So, even if your Global Survey confirms the explosive growth of AI, you should not be celebrating Generative AI’s breakout year and hold off celebrating until someone manages to put this destructive brain-dead genie we’ve unleashed back into the bottle it was released from!

Dear (Software) Vendor: If you Missed the Ten (+ 2 Bonus) Best Practices for Success, Time to Catch Up Now!

  • Part 1 Best Practices #1 to #3
  • Part 2 Best Practices #4 to #7
  • Part 3 Best Practices 8 to 10
  • Part 4 Bonus Best Practice #1
  • Part 5 Bonus Best Practice #2

In twenty years as an independent analyst and consultant, the doctor has never encountered a small/mid-size vendor who wasn’t doing at least one of these, usually there were a couple they weren’t doing, and the lack of these practices (and knowledge) was (and sometimes still is) holding these vendors back. In other words, you definitely should read these. We are only posting these articles once.

DO NOT CONFUSE THE ILLUSION OF UNDERSTANDING WITH ACTUAL UNDERSTANDING!

Because if you do, you will believe AI is Actually Intelligent when, in fact, as we have pointed out again and again and again, it is Artificial Idiocy, and the best modern technology only uses AI for thunking, not thinking, as thinking needs to remain the domain of us humans (before X robs us of our ability to use actual words).

Not only is there no AI, but when you type a command, there isn’t even any understanding by the algorithm of what you are asking for when you type a query into an AI tool. NONE. It’s all based on a statistical algorithm that uses pre-computed similarity probabilities to infer what you are asking. That’s not understanding. Not even close.

The Guardian recently published a long read article on Weizenbaum’s nightmares: how the inventor of the first chatbot turned against AI that anyone who is even mildly contemplating an AI tool needs to read. Slowly and carefully. Three times.

Weizenbaum, who was a mathematician, computer scientist, and a student of psychoanalysis, was one of the founders of modern artificial intelligence who not only invented the first chatbot (Eliza), but also built early (mainframe) computers (back when they used vacuum tubes and took up entire rooms) for the University he was studying at, General Electric, and the Navy. In the 1960s, he was part of Project MAC at MIT, a Pentagon program for “machine aided cognition” that perfected time-sharing, created in-system messaging (like instant messaging or early email), and created new tools for word processing.

He was also one of the first to think about the implications of Artificial Intelligence years, if not decades, before anyone else and one of the founders of computer ethics. He was a genius, and when he said that Artificial Intelligence is an “index of the insanity of our world“, he was totally right — and he was right five decades before AI became the buzz-acronym-du-jour. Few people effectively saw that far ahead in technology, so maybe we should sit back and listen. Carefully.

So please take the time to read Weizenbaum’s nightmares: how the inventor of the first chatbot turned against AI and realize that AI is not the answer. Deterministic algorithms developed by smart people that have studied the problem, tested their assumptions, and been consistently proven reliable are the answer. They may be based on machine learning, but machine learning that is expertly selected, tuned, and monitored by validation code that detects when the algorithm is not performing to expectation and interjects a human into the process. Not a multi-layered pseudo-random statistical algorithm that randomly predicts the next seven days worth of orders, starting on Monday, are 210, 198, 307, 250, 185, 250, and 3095 and thinks everything is A-OK even though the store is closed on Sunday.