Digital Thread is still a Debate in 5th IR?

[@EthicsAndEngineering]

To add to the discussion of data objects moving up the stack, the book "Algorithms to Live By: The Computer Science of Human Decisions" by Tom Griffiths and Brian Christian do a deeper dive into the computer science concepts that this video refers to. I strongly recommend this book as a supplement - everything from game theory, sorting algorithms, what "computational kindness" is, randomization, ML/AI, and more. Hope this helps!

[@4.0Solutions]

Thanks for the recommendations Enri!

[@almukhtar77]

Awesome, thanks now is much clearer

[@4.0Solutions]

Glad it helped!

[@BretWrightSTUFF]

Going to watch again!!

[@4.0Solutions]

That’s what we love to hear!❤️

[@pjgeldenhuis1]

@Walker Reynolds - thanks for another great video and the detailed example at the end. Curious, how would you architect this if you had to send information from the UNS back to the payment system? Also, under what circumstances would you want to create a digital thread vs. directly connecting to the UNS and why?

[@4.0Solutions]

Great questions, PJ. Will try to answer in tomorrow's podcast.

[@mdrashedulislam7037]

lots of love from Bangladesh ❤❤❤❤❤

[@4.0Solutions]

Sending love back to you from Texas ❤️

[@SolidStateSoul]

Great discussion and important topics, appreciate the visual aids.

[@4.0Solutions]

Glad you enjoyed it! Let us know what else you'd like to see us cover.

[@SolidStateSoul]

@@4.0Solutions I'd like to see how you guys would utilize GPT-3.5 learned on a customer specific data lake to perform process oriented analytics. For example, I work for a large industrial OEM that produces PLC driven machines. I'd like to capture data from our field machines into a data lake, which in tandem is modeled via GPT-3.5 for contextualized output. Meaning our engineers could ask, "How could our customers improve OOE for our equipment based in this region using this fuel?" or "Which of our customers' machines' primary drives are closest to a fail condition, who are also customers likely to pay for preventative maintenance?" or "Give me a list of 5 customers who would benefit from an equipment expansion on their site, and provide me scenario specific operational increase ratios." Things like this are possible with a chat AI that is modeled on specific data sets (which would always be increasing in size, to be modeled cyclically). All of this is possible today, but few integrators have a clue how to do it.

[@4.0Solutions]

Terrific suggestion! We've been doing a series of workshops leveraging ChatGPT for development, next one on IIOT solutions. If you'd be interested in co-hosting a workshop or podcast on the subject, email us at [e-mail].

[@JorgenKvinge]

Very nice and illustrative video, another one in the bank for a better understanding of different ways to approach the same problem (and why UNS makes so much sense). Beside the main message of the video, what you talk about in the end regarding license cost; how do you think NLP/CGPT etc will affect the SW market and licensing cost going forward? I guess making long term contracts on licenses now should be avoided, and creating a specialized team internally (for developing what you want) instead of using SI`s could pay off rather quick if done correctly.

[@walkerreynolds973]

Spot on - licensing costs should definitely come down, especially on the maintenance side. That being said - off the shelf platforms + build will become the norm.

[@MichaelRada-INDUSTRY50]

Dear Walker, I am confused because it looks like you forgot our years-old discussion and recording of my talk with your colleague fully dedicated to INDUSTRY 5.0 and principles /I put the link below). I would like to advise you to look for a new term to describe what you do, but for sure INDUSTRY 5.0 is not the right one, because using the term mislead your audience, which is something you are warning about. I hope it would be fine if I share the only existing INDUSTRY 5.0 DEFINITION, which was published already on February 3rd, 2017, and has over 100.000 views on original and mirrored locations

[@kilojoules88]

Isn't the fifth industrial revolution now threatening to upend your mission statement (boosting the middle class by bringing back manufacturing jobs)? If anything, it looks like we will have the AI tools to automate a significant share of jobs over the next few years, and virtually all jobs on a mid- to long-term horizon.

[@walkerreynolds973]

That depends entirely on how AI is leveraged. Jobs are a function of demand, demand a function of value - AI like NLP increases the value of the individual in a market, if used correctly. Economies advance - our advancement has been slowed by limited time, money, human capital - AI speeds up output by 10x - 300x - so, it follows that economic advancement will speed up - rather than the demand for human value drying up.

[@kilojoules88]

@@walkerreynolds973 Sure, in an intermediate stage I agree. What's next though? As soon as we reach full AGI (which may be only a few years out, or longer - experts disagree with each other), jobs will concentrate to very few managerial positions, and then ultimately these will be eliminated as well.

[@walkerreynolds973]

@@kilojoules88 A hyper specialized economy -- we will see a split in 'work' based on EQ and IQ (the ability to get the most out of AI). We will see an explosion in cultural and entertainment 'jobs' and we will see massive growth in industrial support positions. Markets that were priced out and left untapped because of low ROI (poor efficiency) will open up. We have discussed this in depth on a couple of our podcasts -- revenue streams for all workers will come from three places, UBI, Hyper Specialized Professions, Gigs.