Doing Things Backwards
It boggles my mind that our data is out there for all the world to see. The dark web runs amok with all kinds of, well, dark, bad things and our personal data is part of it. Every day data breaches, ransomware attacks, occur — so many there are sites like this to track it. We have a joke where we live about not wanting to be on the cover of the Carmel Pine Cone. Well, let me tell you it’s not so cool to have your brand listed on this tracker either.
A couple weeks ago I read that a lot of organizations impacted by the CrowdStrike so-called “software upgrade” put their back-up plans into place and nada. They didn’t work. You think you can plan for any eventuality but then whammo you are s-o-l and in the news.
Delta joined Southwest Airlines who was carrying the title for catastrophic melt-down award of the year — circa Christmas 2023 — on the list of epic failures. And it’s ironic to me that all this was done by a company who was supposed to be doing good! Backwards, but here we are.
Enter the world of GenAI, data provenance matters; data integrity matters; hallucinations matter; humans in the loop matter. The stakes are only getting higher. My colleague David Worlock mentioned this the other day:
“….a number of pieces that I have noted recently are about the use of the reasoning capacity in Strawberry AI when combined with its autonomous web searching capacity. Strawberry is the codename for developmental work done by Open AI in preparation for the launch of its new Orion environment. It first surfaced in the AI industry newsletters last month, but there is now a torrent of commentary about it. I’ve copied a sample below.
The primary goal of Strawberry is to improve AI’s ability to generate not just answers but to autonomously conduct thorough research. This represents a significant leap towards creating AI models capable of advanced planning and problem-solving, moving closer to human-like reasoning abilities. A key aspect of Strawberry involves “post-training” techniques. Unlike traditional methods that rely solely on initial training with large datasets, Strawberry refines AI models after this initial phase. This specialized approach aims to enable AI to tackle complex tasks over extended periods, a challenge that existing models often struggle with.
If you want more background, then the sources on Perplexity can provide it. “
As we turn more and more over to the machines so that we can all be blessed with a universal income (thank you Sam) we can’t even keep our technology and data secure. Why must humans do things backwards? Memo to Boeing — you do not put people UP in space until you are sure you can get them DOWN. Yes, we are pioneers living on the edge and accidents happen when you pioneer. But leaving people stranded in outer space when you can’t keep planes in the air is an adjacency too far.
My son is trying to unwind Carfax who decided that each time he moved and registered his car in a new state it was a ‘sale.’ He was surprised to find that his car has had four owners even though he’s three of them! It makes no sense. He must now fix the mess they made, and the onus is on the customer to fix it. Backwards is an understatement. Someone actually programmed this logic!
Finally, National Public Data is getting some heat. Egregious mess if there ever was one.
And recently after buying a car, before I knew it, the sales guy had me signed up, signed on, and everything ‘recorded’ up to the manufacturers cloud….so I could drive off the lot. For my safety, I am told. For my insurer, I think. Never in the 30-minute ramp-time to rev-time was I asked for permission about gathering this data, told who owned it, or what would be done with it. It is now up to me to ask, then research, then unwind if I ever can. It’s on me now.
Just like it’s on us to call the credit bureau to freeze credit reports or find a company to get us off the dark web (good luck with that!) or get Carfax to straighten up. Data companies have got to do better even if Delta Airlines and Boeing can’t!
So going forward, what’s an exec to do?
Join us for the Outsell Signature Event, co-produced with JEGI CLARITY where we’ll hear a short video interview with John Marcante. Founder and CEO of Technology Leadership Solutions LLC about Tech Security: Prevention, Resiliency, Recovery — What CEOs and Boards Need to Be Asking. John is an expert who has joined us before and who you will love hearing. He is a wealth of knowledge about how to get this right — and more important, provides essential questions we need to be asking.
He’ll join us by video again in November (Members Only) for an interactive session on these same topics. You won’t want to miss him. Register today; become a member today!