I’m all for something that will help college students plan and implement effective study strategies. I’ve been out of higher education for a few years now but I am aware of learning management systems moving in this direction, albeit more for instructors and advising teams to collect data than to guide students. What you are proposing is more at the level of the personal study assistant. I applaud the effort and direction.
Allow me to offer some critical perspective based on a few things I see and don’t see in this description. The ed-tech space has seen many attempts at, and promises about, personalization that haven't delivered meaningful results. The app appears to be content agnostic, so it is even more important that it gets some fundamentals about learning right, and avoid what is ineffective.
How well is the knowledge base of the model grounded in evidence for how people learn? When I read "learning patterns and preferences," I hear the theory of “learning styles,” a thoroughly debunked and pernicious myth. I looked for, but did not see, evidence-based principles from cognitive science that are applicable in all learning domains: spacing, retrieval, elaboration, generation, interleaving, and dual-coding to name several. Absent these, I am concerned about how effectively this tool will guide students.
We want students to develop accurate metacognition and self-regulation. A study assistant should be grounded in how to give students the right feedback, and guide them to make the right effort, to develop those skills. Effort is the key ingredient. A study assistant that removes cognitive effort may be removing the work that permits growth. I didn’t see the terms “metacognition” or “self-regulation” so these may be a significant gap in the assistant’s knowledge base and how it applies algorithms to giving feedback and direction.
What is “cognitive rhythm?” I am ABD in instructional design and have advocated for evidence-informed practice for many years. No such concept exists in the literature that I know of. Feedback on choosing a study environment could be helpful. Does the app understand that varied context can be helpful for learning? Again, I worry that conventional wisdom may replace more evidence informed and nuanced recommendations.
While this is billed as personal assistant, I have doubts about an algorithm, no matter how sophisticated, being able to ascertain and solve the cognitive or emotional barriers unique to each student. AI, lacking a theory of mind, cannot effectively assess and address conceptual or emotional blocks particular to each student in a given moment. Moreover, a tool that promotes the independence (isolation!?) depicted in your scenario takes humans out of the loop, to the potential detriment of learning. I don’t see it directing students to helpful humans. Be careful not to oversell what technology can do
Perhaps I have misread the proposal or read too much into it. This is how it lands for me. Take it as one potential “customer’s” feedback. Good luck! @benjamin
Thank you, @Benjamin, for taking the time to share such thoughtful and detailed feedback. I really appreciate your insights, especially given your background in instructional design and evidence-informed practices.
To clarify, **ATLAS is a prototype developed during a 3-day hackathon** by an incredibly talented team of developers. While I’m not directly involved in its future development, your points highlight key areas they could consider as they refine the concept.
1) Grounding in Cognitive Science
You’re absolutely right about the importance of evidence-based principles like spacing, retrieval, and interleaving. While these weren’t explicitly incorporated in the prototype, they should absolutely be a foundation for any future work. Avoiding myths like “learning styles” is a great reminder for tools like this.
2) Metacognition and Self-Regulation
I completely agree that helping students develop metacognitive skills and self-regulation is crucial. This wasn’t fully addressed in the prototype, but your feedback makes it clear how important it is for a tool like this to encourage effort, reflection, and growth.
3) Cognitive Rhythms and Evidence-Based Practice
The term “cognitive rhythms” was meant to describe individual patterns in focus and energy, but I see how that might be unclear. Revisiting the terminology and grounding it in established research is definitely something for the team to consider.
4) Humans in the Loop
I love your point about keeping human support in the loop. While ATLAS was designed to complement existing resources, not replace them, building connections to advisors, instructors, or peers would make it even more impactful.
Thank you again for your thoughtful critique and encouragement. While I’m not responsible for ATLAS’s future development, I’ll be sure to share your feedback with the team—it’s incredibly valuable and will give them a lot to think about as they move forward :)
Agreed! Not an expert by any means, but I always had a problem with well intentioned people thinking that GenAI could accurately predict and interpret how different students learn.
You explained that clearly in your comment.
Another thing I would explore is how many of these ed tech apps don’t have people who are not neurotypical in mind.
For example, a LLM like ChatGPT might be good for someone that has ADHD, but not someone that has severe autism.
henomenal support for students. As someone who is a non techie/beginner level, who loves all things technology and sees so many possibilities for AI /Gen AI/AI Agents to make our lives better. I am wondering if there are resources for people that are basically 'plug and play'?
This is now written as a code tutorial. All you need to do is run all the cells by pressing the "play" button, and it should work. It’s not entirely a plug-and-play product, but it’s fairly straightforward. Let me know if it works for you!
What a fantastic exploration of using AI agents to support students' academic journeys! Your ATLAS system presents an innovative approach to addressing the complex challenges modern students face. I'd love to share how we're thinking about complementary dimensions of this important challenge.
While ATLAS offers powerful AI agents for individual academic support, we're exploring how Omnimity could take this concept further by treating all elements of online academic discourse - from class discussions and student interactions to assignment feedback and peer reviews - as self-organizing agents that naturally surface the most relevant educational content and connections. Rather than just managing tasks and schedules, this approach would recognize and harness the intrinsic value emerging from classroom dialogue, allowing credentials to form organically based on students' demonstrated knowledge and contributions.
Omnimity allows classroom discussions to automatically connect to real-world applications, linking students with relevant industry experts, groups or communities, or practical challenges or opportunities that match their emerging expertise. As students engage in course discussions and projects, their contributions would generate dynamic, verifiable credentials that extend beyond traditional grades - credentials that reflect their actual capabilities and continue evolving throughout their careers. Universities could maintain complete control over their data while still participating in this broader knowledge ecosystem, creating seamless pathways between academic learning and professional opportunities. This shifts education from a time-bounded experience to a lifelong journey of growth and contribution, where the boundaries between classroom and real-world learning naturally dissolve through the power of self-organization.
This is all part of what we are developing with Omnimity and an essential subject in my book Nature's Algorithm, which provides the theory behind this design. It doesn't just apply to education, but ever segment of society.
Thanks for sharing your thoughts! The idea of self-organizing agents and evolving credentials is really interesting and seems to take a broader approach than what ATLAS focuses on. ATLAS is more about tackling the immediate challenges students face with personalized academic support, but I can see how Omnimity could complement that by connecting learning to real-world applications. I’d be curious to hear more about how this works in practice and what challenges you’ve faced with implementation. Thanks for contributing to the discussion!
In practice, Omnimity analyzes classroom discussions and online dialogue to recognize subject matter expertise and form meaningful connections organically. The biggest implementation challenge is working within existing educational platforms, which are often restrictive and centralized. We're addressing this by developing Omnimity as a complementary layer that lets institutions maintain control of their data (autonomy) while enriching their current systems with self-organizing capabilities that connect classroom learning to real-world opportunities. We're also establishing a non-profit organization to build relationships with educational institutions and support implementation in a way that prioritizes student and institutional needs over commercial interests.
Are you familiar with any online education platforms or resources that might be open as an early adapter to a technology like this?
I’m all for something that will help college students plan and implement effective study strategies. I’ve been out of higher education for a few years now but I am aware of learning management systems moving in this direction, albeit more for instructors and advising teams to collect data than to guide students. What you are proposing is more at the level of the personal study assistant. I applaud the effort and direction.
Allow me to offer some critical perspective based on a few things I see and don’t see in this description. The ed-tech space has seen many attempts at, and promises about, personalization that haven't delivered meaningful results. The app appears to be content agnostic, so it is even more important that it gets some fundamentals about learning right, and avoid what is ineffective.
How well is the knowledge base of the model grounded in evidence for how people learn? When I read "learning patterns and preferences," I hear the theory of “learning styles,” a thoroughly debunked and pernicious myth. I looked for, but did not see, evidence-based principles from cognitive science that are applicable in all learning domains: spacing, retrieval, elaboration, generation, interleaving, and dual-coding to name several. Absent these, I am concerned about how effectively this tool will guide students.
We want students to develop accurate metacognition and self-regulation. A study assistant should be grounded in how to give students the right feedback, and guide them to make the right effort, to develop those skills. Effort is the key ingredient. A study assistant that removes cognitive effort may be removing the work that permits growth. I didn’t see the terms “metacognition” or “self-regulation” so these may be a significant gap in the assistant’s knowledge base and how it applies algorithms to giving feedback and direction.
What is “cognitive rhythm?” I am ABD in instructional design and have advocated for evidence-informed practice for many years. No such concept exists in the literature that I know of. Feedback on choosing a study environment could be helpful. Does the app understand that varied context can be helpful for learning? Again, I worry that conventional wisdom may replace more evidence informed and nuanced recommendations.
While this is billed as personal assistant, I have doubts about an algorithm, no matter how sophisticated, being able to ascertain and solve the cognitive or emotional barriers unique to each student. AI, lacking a theory of mind, cannot effectively assess and address conceptual or emotional blocks particular to each student in a given moment. Moreover, a tool that promotes the independence (isolation!?) depicted in your scenario takes humans out of the loop, to the potential detriment of learning. I don’t see it directing students to helpful humans. Be careful not to oversell what technology can do
Perhaps I have misread the proposal or read too much into it. This is how it lands for me. Take it as one potential “customer’s” feedback. Good luck! @benjamin
Thank you, @Benjamin, for taking the time to share such thoughtful and detailed feedback. I really appreciate your insights, especially given your background in instructional design and evidence-informed practices.
To clarify, **ATLAS is a prototype developed during a 3-day hackathon** by an incredibly talented team of developers. While I’m not directly involved in its future development, your points highlight key areas they could consider as they refine the concept.
1) Grounding in Cognitive Science
You’re absolutely right about the importance of evidence-based principles like spacing, retrieval, and interleaving. While these weren’t explicitly incorporated in the prototype, they should absolutely be a foundation for any future work. Avoiding myths like “learning styles” is a great reminder for tools like this.
2) Metacognition and Self-Regulation
I completely agree that helping students develop metacognitive skills and self-regulation is crucial. This wasn’t fully addressed in the prototype, but your feedback makes it clear how important it is for a tool like this to encourage effort, reflection, and growth.
3) Cognitive Rhythms and Evidence-Based Practice
The term “cognitive rhythms” was meant to describe individual patterns in focus and energy, but I see how that might be unclear. Revisiting the terminology and grounding it in established research is definitely something for the team to consider.
4) Humans in the Loop
I love your point about keeping human support in the loop. While ATLAS was designed to complement existing resources, not replace them, building connections to advisors, instructors, or peers would make it even more impactful.
Thank you again for your thoughtful critique and encouragement. While I’m not responsible for ATLAS’s future development, I’ll be sure to share your feedback with the team—it’s incredibly valuable and will give them a lot to think about as they move forward :)
Agreed! Not an expert by any means, but I always had a problem with well intentioned people thinking that GenAI could accurately predict and interpret how different students learn.
You explained that clearly in your comment.
Another thing I would explore is how many of these ed tech apps don’t have people who are not neurotypical in mind.
For example, a LLM like ChatGPT might be good for someone that has ADHD, but not someone that has severe autism.
those are some excellent insights. I agree
henomenal support for students. As someone who is a non techie/beginner level, who loves all things technology and sees so many possibilities for AI /Gen AI/AI Agents to make our lives better. I am wondering if there are resources for people that are basically 'plug and play'?
Thanks!
This is now written as a code tutorial. All you need to do is run all the cells by pressing the "play" button, and it should work. It’s not entirely a plug-and-play product, but it’s fairly straightforward. Let me know if it works for you!
Hello DiamantAI,
What a fantastic exploration of using AI agents to support students' academic journeys! Your ATLAS system presents an innovative approach to addressing the complex challenges modern students face. I'd love to share how we're thinking about complementary dimensions of this important challenge.
While ATLAS offers powerful AI agents for individual academic support, we're exploring how Omnimity could take this concept further by treating all elements of online academic discourse - from class discussions and student interactions to assignment feedback and peer reviews - as self-organizing agents that naturally surface the most relevant educational content and connections. Rather than just managing tasks and schedules, this approach would recognize and harness the intrinsic value emerging from classroom dialogue, allowing credentials to form organically based on students' demonstrated knowledge and contributions.
Omnimity allows classroom discussions to automatically connect to real-world applications, linking students with relevant industry experts, groups or communities, or practical challenges or opportunities that match their emerging expertise. As students engage in course discussions and projects, their contributions would generate dynamic, verifiable credentials that extend beyond traditional grades - credentials that reflect their actual capabilities and continue evolving throughout their careers. Universities could maintain complete control over their data while still participating in this broader knowledge ecosystem, creating seamless pathways between academic learning and professional opportunities. This shifts education from a time-bounded experience to a lifelong journey of growth and contribution, where the boundaries between classroom and real-world learning naturally dissolve through the power of self-organization.
This is all part of what we are developing with Omnimity and an essential subject in my book Nature's Algorithm, which provides the theory behind this design. It doesn't just apply to education, but ever segment of society.
Would love to know your thoughts.
Thanks for sharing your thoughts! The idea of self-organizing agents and evolving credentials is really interesting and seems to take a broader approach than what ATLAS focuses on. ATLAS is more about tackling the immediate challenges students face with personalized academic support, but I can see how Omnimity could complement that by connecting learning to real-world applications. I’d be curious to hear more about how this works in practice and what challenges you’ve faced with implementation. Thanks for contributing to the discussion!
In practice, Omnimity analyzes classroom discussions and online dialogue to recognize subject matter expertise and form meaningful connections organically. The biggest implementation challenge is working within existing educational platforms, which are often restrictive and centralized. We're addressing this by developing Omnimity as a complementary layer that lets institutions maintain control of their data (autonomy) while enriching their current systems with self-organizing capabilities that connect classroom learning to real-world opportunities. We're also establishing a non-profit organization to build relationships with educational institutions and support implementation in a way that prioritizes student and institutional needs over commercial interests.
Are you familiar with any online education platforms or resources that might be open as an early adapter to a technology like this?
not at the moment, but will let you know as soon as I hear something :)