I don't mean to be an extremist, but this policy creates a system where academic work is no longer judged solely on its rigor and merit, but on its political alignment. The policy also weaponizes the very digital tools we use for learning and turns them into instruments of surveillance and compliance.
As an educator who incorporated AI into my classrooms, I encourage students to use it for any and all assignments as a way to think critically about any problem before them.
I see it as effectively silencing any inquiry that challenges a specific, state-approved definition of "objective truth."
Thank you for the reply Kamil. I’ve read 90% of your posts. Really appreciate them. I’ve been trying to twist my employer to cover a subscription too. To your point. I do agree.
great summary, here are my takes related to Empowering American Workers in the Age of AI
The Trump Administration is doubling down on a worker-first AI agenda, recognizing that while AI will accelerate productivity and create new industries, it will also redefine how work is done. Here’s what’s new:
💼 Key Actions to Prepare the U.S. Workforce:
🧠 AI Literacy at Scale: Federal education and workforce programs—including CTE, apprenticeships, and workforce retraining—will prioritize AI skill development.
🏛️ Tax-Free Upskilling: Treasury will clarify how AI education can qualify as employer-provided, tax-free educational assistance under IRC Section 132.
📊 Labor Market Intelligence: BLS, Census, and BEA will track AI’s impact on job creation, displacement, and wages.
🧪 AI Workforce Research Hub: A new DOL-led entity will lead scenario planning and deliver actionable insights for policymakers.
🚀 Rapid Retraining Pilots: Programs will be launched to proactively retrain and upskill Americans at risk of AI-related job displacement.
Good. But the ‘shift’ part is a little unclear. What actually represents a change from previous administrations?
…also not clear what is policy (which actually affects the real world) versus an outlook/wishlist. Ok, so agencies can’t buy woke AI, and regulatory agencies will be (stay?) hands off at the federal level…
But let’s be clear: so far the Trump Administration is doing the exact *opposite* of what’s called for re:grid/electricity. Whether you’re a wind/solar fan or not, it is the fastest growing electrical generation segment in the US. The Big Beautiful Bill will unequivocally put the brakes on new construction and give away the global lead to China et al over thr next few years… and to my knowledge the new law doesn’t replace renewable subsidies with subsidies on other electricity sources (or grid expansion incentives etc). Would love to see Trump reform the Nuclear Regulatory Commission, for example…make it easier to build SMRs, easier to expand/extend the life of current BWR/PWR fleet.
Should be incredibly interesting to follow, watching the innovation and the best ways to utilize this new technology. I don’t work so I’m more on the sidelines watching to see how close my thoughts on how it will work successfully is close to what is being developed and used. I’m on my computer once to twice a week so trying to really dive in like most of your followers, isn’t really feasible for me. I’m waiting and watching how companies are at meeting customers where they are, offering AI as a service for users. How they manage the customer personal information and data locally I think will make the difference in if they are successful. Multiple different AI systems might act as a security check on the local network even if all of them are created from open source. How the system uses the different AI customized and offered as a service with all other AI services being utilized on a local network could create its own cryptography. If each are seen as reinforcement of permissions by knowing how many other sources of personal information validated with confirmation by each different link of information from who has issued it and kept together in the local wallet. The more sources of validated points of information strengthens the understanding from other sources that it’s a valid system. So if my state id has been approved through the DMV, my different cards or payment accounts are validated through their individual companies that issued them, software or AI service certificates of valid ownership or permission for use from this account works as a green check point in the system. They can’t access each other or see the data but they will know how many data sources that have validation and locked in the system reinforces a more complex security lock in the system when the number of accounts matches the number of accounts the account should have. If a connection is broken and cannot not be certain that it’s a valid point of source data information used for the handshake validation then the other points are frozen and are read as a breach in the system with possible compromise in the proper account permissions that manages what can be accessed in the system to use it. So if an AI application goes rogue and tries to change anything outside of what they have been given permission to use for the service. So different AI services would have stricter guidelines and guardrails they would have to stay within. That would possibly act as a vault security in the system the assures that they won’t tamper with any other identity validation data or permissions. I had a hard time trying to put what I have in mind into words, I really hope this makes sense. If it does… then you understand what I’m hoping to see as a personal customer user to have available.
Anyway. My vision as a user I think needs to be narrowed down to the information accessed for the AI service to be able to fully benefit from the service without having to sacrifice space, speed and power necessary. As a personal user and not a business I don’t have the hardware necessary to run full systems locally. I also don’t want personal data used in the service to be accessed off of my system. The more I have to share outside of my system the greater chances my identity and information will be compromised. To actually be useful to individual customers and a successful service, it will have access to medical records, care information, contracts with insurance companies both property and health, information that is used to manage my home like bills, inventory of different items and products, knowledge of my assets, etc. As tempting as that data is to companies, they need to come to the understanding of the damage that could be done by their service if compromised. It’s not a controllable environment like a personal computer being accessed through a data connection. The way it will need to blend with the data given to work outside of the traditional file/folder system to a more tagged and smart organization of information by given parameters of exactly what information is needed to accomplish what is needed and dynamic enough that it is able to conform to similar tasks but different information even slightly. It will need to have it catalogued in a manner that it needs to use the information properly in the service without having to ramp up hardware needed to process everything each time it’s used or maintained off the system. Even the small continuous connection with the company system will cause problems with resources needed to maintain it. It just won’t be efficient and scalable for the company or customers with big models that most are building to be useful in a smooth experience and a profitable product for the company to offer an maintain, setting it up for failure. Businesses can stumble and modify but failure for the customer can cause a loss in confidence or the desire to constantly change how something is done locally to meet expectations to see the value in services or different companies. Too many businesses are rushing to be king of the AI models with profits that allow the company to grow beyond a small independent company.
I don't mean to be an extremist, but this policy creates a system where academic work is no longer judged solely on its rigor and merit, but on its political alignment. The policy also weaponizes the very digital tools we use for learning and turns them into instruments of surveillance and compliance.
As an educator who incorporated AI into my classrooms, I encourage students to use it for any and all assignments as a way to think critically about any problem before them.
I see it as effectively silencing any inquiry that challenges a specific, state-approved definition of "objective truth."
I think it's going to be a challenge to have an unbiased system to begin with.
Thank you for the reply Kamil. I’ve read 90% of your posts. Really appreciate them. I’ve been trying to twist my employer to cover a subscription too. To your point. I do agree.
HR usually has a L&D budget that can cover it :)
Dm me if you need help and thank you for reading Dave
Buckle up, Buttercup! This is going to be an interesting ride! Thanks for your insights!
great summary, here are my takes related to Empowering American Workers in the Age of AI
The Trump Administration is doubling down on a worker-first AI agenda, recognizing that while AI will accelerate productivity and create new industries, it will also redefine how work is done. Here’s what’s new:
💼 Key Actions to Prepare the U.S. Workforce:
🧠 AI Literacy at Scale: Federal education and workforce programs—including CTE, apprenticeships, and workforce retraining—will prioritize AI skill development.
🏛️ Tax-Free Upskilling: Treasury will clarify how AI education can qualify as employer-provided, tax-free educational assistance under IRC Section 132.
📊 Labor Market Intelligence: BLS, Census, and BEA will track AI’s impact on job creation, displacement, and wages.
🧪 AI Workforce Research Hub: A new DOL-led entity will lead scenario planning and deliver actionable insights for policymakers.
🚀 Rapid Retraining Pilots: Programs will be launched to proactively retrain and upskill Americans at risk of AI-related job displacement.
Re: Rapid Retraining Pilots
Start with Air India…?
Anything but boring in the next couple of years - indeed.
Good. But the ‘shift’ part is a little unclear. What actually represents a change from previous administrations?
…also not clear what is policy (which actually affects the real world) versus an outlook/wishlist. Ok, so agencies can’t buy woke AI, and regulatory agencies will be (stay?) hands off at the federal level…
But let’s be clear: so far the Trump Administration is doing the exact *opposite* of what’s called for re:grid/electricity. Whether you’re a wind/solar fan or not, it is the fastest growing electrical generation segment in the US. The Big Beautiful Bill will unequivocally put the brakes on new construction and give away the global lead to China et al over thr next few years… and to my knowledge the new law doesn’t replace renewable subsidies with subsidies on other electricity sources (or grid expansion incentives etc). Would love to see Trump reform the Nuclear Regulatory Commission, for example…make it easier to build SMRs, easier to expand/extend the life of current BWR/PWR fleet.
Should be incredibly interesting to follow, watching the innovation and the best ways to utilize this new technology. I don’t work so I’m more on the sidelines watching to see how close my thoughts on how it will work successfully is close to what is being developed and used. I’m on my computer once to twice a week so trying to really dive in like most of your followers, isn’t really feasible for me. I’m waiting and watching how companies are at meeting customers where they are, offering AI as a service for users. How they manage the customer personal information and data locally I think will make the difference in if they are successful. Multiple different AI systems might act as a security check on the local network even if all of them are created from open source. How the system uses the different AI customized and offered as a service with all other AI services being utilized on a local network could create its own cryptography. If each are seen as reinforcement of permissions by knowing how many other sources of personal information validated with confirmation by each different link of information from who has issued it and kept together in the local wallet. The more sources of validated points of information strengthens the understanding from other sources that it’s a valid system. So if my state id has been approved through the DMV, my different cards or payment accounts are validated through their individual companies that issued them, software or AI service certificates of valid ownership or permission for use from this account works as a green check point in the system. They can’t access each other or see the data but they will know how many data sources that have validation and locked in the system reinforces a more complex security lock in the system when the number of accounts matches the number of accounts the account should have. If a connection is broken and cannot not be certain that it’s a valid point of source data information used for the handshake validation then the other points are frozen and are read as a breach in the system with possible compromise in the proper account permissions that manages what can be accessed in the system to use it. So if an AI application goes rogue and tries to change anything outside of what they have been given permission to use for the service. So different AI services would have stricter guidelines and guardrails they would have to stay within. That would possibly act as a vault security in the system the assures that they won’t tamper with any other identity validation data or permissions. I had a hard time trying to put what I have in mind into words, I really hope this makes sense. If it does… then you understand what I’m hoping to see as a personal customer user to have available.
Anyway. My vision as a user I think needs to be narrowed down to the information accessed for the AI service to be able to fully benefit from the service without having to sacrifice space, speed and power necessary. As a personal user and not a business I don’t have the hardware necessary to run full systems locally. I also don’t want personal data used in the service to be accessed off of my system. The more I have to share outside of my system the greater chances my identity and information will be compromised. To actually be useful to individual customers and a successful service, it will have access to medical records, care information, contracts with insurance companies both property and health, information that is used to manage my home like bills, inventory of different items and products, knowledge of my assets, etc. As tempting as that data is to companies, they need to come to the understanding of the damage that could be done by their service if compromised. It’s not a controllable environment like a personal computer being accessed through a data connection. The way it will need to blend with the data given to work outside of the traditional file/folder system to a more tagged and smart organization of information by given parameters of exactly what information is needed to accomplish what is needed and dynamic enough that it is able to conform to similar tasks but different information even slightly. It will need to have it catalogued in a manner that it needs to use the information properly in the service without having to ramp up hardware needed to process everything each time it’s used or maintained off the system. Even the small continuous connection with the company system will cause problems with resources needed to maintain it. It just won’t be efficient and scalable for the company or customers with big models that most are building to be useful in a smooth experience and a profitable product for the company to offer an maintain, setting it up for failure. Businesses can stumble and modify but failure for the customer can cause a loss in confidence or the desire to constantly change how something is done locally to meet expectations to see the value in services or different companies. Too many businesses are rushing to be king of the AI models with profits that allow the company to grow beyond a small independent company.