In the gleaming conference rooms of Silicon Valley’s most prestigious AI companies, a young Indian engineer named Aravind Srinivas was watching the future unfold. It was 2021, and the 27-year-old PhD from UC Berkeley had just spent years rotating through the holy trinity of artificial intelligence research: OpenAI, Google DeepMind, and back to OpenAI again. He had seen the inner workings of ChatGPT before the world knew its name, witnessed the transformer revolution firsthand, and absorbed the cutting-edge research that would soon reshape human knowledge.
Aravind Srinivas, Co-Founder & CEO of Perplexity. Photo by Kimberly White Getty Images for TechCrunch.
For this survey, I asked seven chatbots — Gemini, DeepSeek, You.com, ChatGPT, Copilot, Perplexity, and Claude — to rank the same seven chatbots, including their own. I weighted each 1st selection as 7 points, each 2nd as 6, etc. Each chatbot received 7 rankings.
The runaway favorite for academic papers was Perplexity with 47 points (six 7s and one 5). The second and third were Claude with 33 and Gemini with 31. Here’s the complete rankings: (1st) Perplexity 47, (2nd) Claude 33, (3rd) Gemini 31, (4th) Copilot 23, (4th) You.com 23, (5th) ChatGPT 21, (6th) DeepSeek 18.
Introduction: These reviews collectively indicate that while ChatGPT-5 represents a significant technical advancement—especially in reasoning, accuracy, and research support—it also introduces new complexities, retains some persistent limitations, and raises important questions about user experience, trust, and ethical deployment. -You.com
The crime & method. Reported in August 2025, an elderly couple received a call from someone who sounded like their grandson, claiming he’d been in a serious crash and jailed for DUI after hitting a pregnant woman. A supposed “official” then demanded $9,000 in bail, arranged an Uber/Lyft pickup to the bank, and later pressed for another $9,000. The station noted the use of AI to mimic the grandson’s voice and the practice of scraping social media to generate a convincing clone. WGAL
1. OpenAI Launches GPT-5: A New Era in AI Reasoning
The unveiling of GPT-5 by OpenAI on August 7, 2025, set a new benchmark for generative AI. GPT-5 introduces a step-change in technical capabilities, especially in reasoning and multi-modal understanding. Its ability to integrate text, image, and voice processing in a unified model marks a significant evolution from previous generations. Benchmarks suggest a 40% improvement over GPT-4 in handling complex reasoning tasks, from scientific problem-solving to autonomous coding and data analysis. champaignmagazine
The crime. In January 2024, a finance staffer in the Hong Kong office of Arup, the global engineering group, received what looked like a legitimate email and invitation to a video meeting with the firm’s UK-based chief financial officer and several colleagues. On the call, everyone appeared and sounded exactly like familiar executives—except they weren’t.
Introduction: In this follow-up to Oregon Trail: Where Two Cultures Collaborate, I worked with Gemini* to detail how the Fletchers (a fictional family) settled in Tualatin Valley, which is in the Willamette Valley, Oregon. -js
Oregon City on the Willamette River in the 1850s, by John Mix Stanley.
Traveling approximately 15 miles per day on average, the Fletchers’ 2,000-mile journey from Independence, Missouri, took approximately five months to complete. When they finally reached Oregon, their primary destination was the Willamette Valley.1 In 1850, the Oregon Territory comprised the entirety of the modern states of Oregon, Washington, and Idaho as well as portions of present-day Montana and Wyoming that lie west of the Continental Divide. The official end of the trail was Oregon City, located at the northern end of the valley, and many people initially gathered there.2
In the world of e-commerce, we talk about “marketplaces,” “platforms,” and “fulfillment networks.” But these words miss the bigger picture: online shopping isn’t just about finding products — it’s a tightly connected loop of searching, purchasing, and delivering.
That’s why I’m introducing SBS™ — Shop · Buy · Ship. It’s a simple but powerful framework for understanding the entire commerce cycle, whether it’s controlled by one company or spread across many.
What Is SBS?
SBS™ (Shop · Buy · Ship) is the integrated process that defines how consumers discover products, pay for them, and receive them.
Introduction: Yesterday, Claude and I collaborated* on a short story, told by Claude from the perspective of a twelve-year-old boy who is traveling the Oregon Trail in a wagon train in 1848. -js
My name’s Thomas Fletcher, but most folks just call me Tommy. Pa says I’m small for twelve, but I reckon I’m doing a man’s work out here, so that’s got to count for something.
We been on this trail near two months now, left Independence back in April when the grass was just starting to green up good for the oxen – though we got mules, not oxen. Pa says mules are steadier, don’t spook as easy. Every morning I’m up before dawn helping Pa check the harnesses and loading our gear back in the wagon. Ma and my sister Sarah, she’s fourteen, they pack up the cooking things and whatever we pulled out for sleeping.
Introduction: I asked Claude to join me in exploring the idea of a smart stadium (s-stadium) for football (American) and soccer that is designed, from the ground up, to maximize the use of AI by coaches and staff to make on-the-fly decisions under game conditions. Realizing that the topic is complex and requires a number of subtopics, I decided that, instead of a single monstrous prompt, a systematic scaffolding of sub-prompts* might be the best approach. The result is this article, which is presented as an informal white paper, with the working title: “The S-Stadium: Designing the AI-Integrated Sports Venue of the Future.” The paper comprises five parts: Part I: Coaching Intelligence Feeds – The Foundation of S-Stadium Design. Part II: Architectural, Engineering, and Technological Systems. Part III: Mother – The S-Stadium AI Orchestration System. Part IV: ARIA – Autonomous Recursive Intelligence Architecture. Part V: Feasibility Analysis – The Reality Check. Epilogue: The Catalyst Factor. -js
The S-Stadium: Designing the AI-Integrated Sports Venue of the Future
An Informal White Paper on Ground-Up Smart Stadium Architecture
Executive Summary
The S-Stadium (Smart Stadium) represents a paradigm shift in sports venue design, where artificial intelligence systems are not retrofitted additions but foundational elements that drive architectural and technological decisions from conception. This paper outlines the tactical video intelligence feeds that would define coaching capabilities in such a venue.
In this 10-minute video posted on YouTube (on 8 June 2025 by the University of Toronto), Ilya Sutskever, co-founder and former chief scientist at OpenAI, speaks at the University of Toronto’s Spring Convocation (6 June 2025). His message: AI is rapidly creating a future that’s “really unprecedented and really extreme,” and the best way to prepare for it is to experience it firsthand, with our own senses, to get a feel for its potential to enhance our ability to learn and perform and to begin to grasp what that means for our future.
Sutskever’s remarks about AI begin at the 3:35 mark of the video below.
Introduction: AI in the field of cyberjustice began to take off only a year ago, in 2024, so it’s a relatively new specialty. In this article, Gemini and I discuss the criminal and justice forces that are defining this field. -js
JS: Gemini, is it possible for criminals to use AI to profit from crimes with a good chance of avoiding arrest?
Gemini: Yes, it’s possible for criminals to use AI to profit from crimes with a good chance of avoiding arrest.1 AI significantly enhances criminals’ capabilities by automating and scaling illegal activities, making them more sophisticated and harder to detect.2 It lowers the barrier to entry for many types of crime, allowing individuals with limited technical skills to execute complex schemes.3
Introduction: In this conversation, ChatGPT and I discuss the idea of extending AI into prisons to boost their rehabilitation mission. Apparently, scattered attempts have been made in limited programs, but there are no clear prototypes for the comprehensive model that we envision. Our opinion is that the benefits, in terms of rehabilitation and compassion, far outweigh the costs and effort. -js
JS: Hi ChatGPT. Let’s dive into the current or potential uses of AI/chatbots in the treatment and rehabilitation of people behind bars or at risk of being there. Is this idea currently being applied or explored in justice systems in the US and world? It seems like a natural trajectory for treating those struggling with what appears to be self-control. If it’s not being considered, what are your thought on its potential? Advantages? Obstacles and possible solutions?
ChatGPT: Hi James! This is an incredibly important and emerging area, and you’re absolutely right: the integration of AI/chatbots into criminal justice reform—especially in treatment and rehabilitation—feels like a natural and overdue step. While it’s still early days, there are a few promising initiatives underway, and the potential is significant.
Introduction: In this conversation, ChatGPT and I are taking a fifteen-year journey into the future, to the year 2040. He’ll be adopting the persona of ʻIlima, a senior at the University of Hawaiʻi at Mānoa, who will be telling us about the University’s huli, or transformation, and how she’s preparing for a career that will secure her financial future as well as her cultural heritage. She’ll also be introducing us to classmates and alumni who are traveling similar but different paths. The focus is on Hawaii, but comparable changes have been occurring in higher ed throughout the US and the world. -js
JS to ChatGPT: Let’s dive into your Model College of 2040 idea. Set this discussion in 2040, 15 years from now. You’ll be the persona, ʻIlima, a female and a senior at the University of Hawaiʻi at Mānoa. You’re part-Hawaiian and part-Asian. UHM, like other colleges across the country and much of the world, has succeeded in radically altering the structure and purpose of college to align programs of study and internships with real-world jobs that provide excellent benefits and salary trajectories that grow from satisfactory to good with experience and performance.*
Introduction:ChatGPT and I looked at the numbers and percentages of four generations (Boomers, Gen X, Millennials, Gen Z) of college graduates and their relative success in the job market. Join us in our conversation as we dive ever deeper into the data and discover a growing crisis. The outlook appears dire, but it’s fixable. It’s a DIY project, and colleges and universities are actually in an excellent time and place to reverse the trend. -js
JS: Can you dig deep for data on the number of college grads who completed their degrees in spring 2024 who are now, a little over a year later, working in jobs that are paying what’s considered a “good” salary? Also, set a figure for what’s considered “good.” How many students graduated in spring 2024? What percentage is earning a “good” salary? Please append an annotated list of references, in APA style, that informed your responses.
ChatGPT: Here’s a detailed summary based on the latest available data:
🎓 1. Estimated Number of Spring 2024 College Graduates (Bachelor’s)
Introduction: My conversation with Gemini began with STEM and moved to: prestige vs non-prestige colleges, non-STEM majors, medicine/nursing, business, soft skills, and transparency of post-graduation salary statistics. We managed to clarify some of the nuances and issues but hit a wall with the issue of greater transparency. -js
JS: Gemini, can you review available data to determine, with some degree of reliability, which colleges are turning out, in proportion to their student population, the greatest number of graduates who are being employed after graduation at the highest salary levels? Let’s limit this survey to US colleges. I realize data may be scarce for some or many colleges, so let’s look at institutions that have data. We can also assume that if the number and prestige of colleges surveyed is low, our generalizability may be compromised.
Introduction: My conversation with ChatGPT began with AI in music and branched into film, dance, animation, and, eventually, the speed of change that’s bypassing most schools and colleges. The early adopters are primarily the young and artists who are learning and practicing their skills independently, creating AI-assisted works with their laptops and personal links to the web. Some institutions are integrating AI into their fine arts programs, but they are the exception. -js
JS: What are some of the most popular ways that musicians, in one or more genres, are using the latest in AI technology? Can you identify some of the major pioneers and how they’re using AI? Are audiences attracted to AI-assisted music?
Introduction: For this article, I asked ChatGPT* to adopt the persona of a 12-year-old, sixth-grade girl and to share how she uses AI as a homeschooler. ChatGPT chose the name “Emi” and Honolulu for the setting. The purpose of this article is to show how AI is being used or could be used in homeschooling. Far from being isolating and monotonous, homeschooling could be exciting and dynamic. The images were generated in ChatBox. -js
Introduction: Kai is a character I created with Perplexity.* He’s a college freshman at Oahu’s Leeward Community College taking English 100, a required course, and tells us, in his own words, what he goes through to complete his first paper with the help of ChatGPT. In the process, he learns there’s a way to go about writing a paper, and it works. I created the images with ChatBox. -js
Introduction: I collaborated with ChatGPT for this report on the status of state, community, school, and college libraries. The conversation, below, took us from libraries reinventing themselves to all‑digital public library systems, from ROI to who uses the library, and from digital vs. traditional libraries to liberal vs. conservative opinions on libraries. -js
Laura Hosman, co-founder and co-director of SolarSPELL
Introduction: I broached the topic for this article — the status of state, community, school, and college libraries — in a prompt* presented to a number of chatbots, and most balked and responded with sketchy outlines. The exception was Gemini, who took the time to deep dive into this complex and, arguably, sensitive topic. The report is published in full below with zero editing (except for the size reduction of main heads). The superscript numbers in Gemini’s on-screen report refer to paragraph-end links that didn’t transfer to ETC Journal. Thus, they don’t link to specific references. -js
Kate Zwaard, formerly of the Library of Congress (LC) and now Executive Director of the Coalition for Networked Information (CNI)
Introduction: I asked Gemini for her opinion on the best free soundtrack creators for non-commercial YouTube videos, and she recommended five. Iʻll be testing them out in the coming weeks and posting updates within this article. If you have any suggestions or comments re these choices, please post them in the comments section attached to this article. Her comments include a very useful explanation of the “free” terms for each choice. -js
Introduction: I collaborated with Gemini on this report. Our search was for a free AI online science lab for secondary students that doesn’t require a physical lab or classroom and can be accessed in a classroom setting or independently outside of school via laptop or desktop over the internet. The LabXchange and PhET Interactive Simulations combo seems to be a good choice. -js
Gemini:LabXchange and PhET Interactive Simulations have joined forces to offer AI-powered hands-on science labs for secondary students. Both platforms, working together, are highly regarded by educators, students, and parents. They are free to use, accessible on laptops and desktops with an internet connection, and provide a way to track student progress over time.
Introduction: For this article, I collaborated with Claude (Sonnet 4). The topic was the probability of humans and dogs someday conversing much as we do with family and friends. My instinct was to request a narrative that might make for a more natural and enjoyable read, and Claude delivered. However, to strengthen the credibility of the narrative, I asked him to append a list of references. He replied that references would require a more academic style. I asked him to proceed and that I would choose one for publication. However, after reading the second version, I couldnʻt decide. They were both effective in their own way. I decided to publish both in this article: the narrative version, first, followed by the academic. Reminder: Claude’s responses come with the usual caveat that applies to all chatbots: Claude can make mistakes, thus, please double-check cited sources. -js
Introduction: I collaborated with Claude, Gemini, Copilot, and Perplexity on this report. The purpose of government is to make its services more accessible and responsive, and the hope is that AI can help to achieve this purpose. -js
Combined responses from Claude, Gemini, Copilot, and Perplexity
Why California Leads in State AI Innovation
Gavin Newsom
Comprehensive AI Strategy: California partnered with NVIDIA to launch a first-of-its-kind AI collaboration in August 2024, signed by Governor Gavin Newsom and NVIDIA founder & CEO Jensen Huang, aimed at training students, educators and workers; supporting job creation and promoting innovation; and using AI to solve problems. This represents one of the most significant public-private AI partnerships at the state level. California also collaborates with companies like Microsoft and Google to test AI tools for data analysis, fraud detection, and predictive modeling in areas like wildfire response and unemployment insurance.
Educational Technology and Change Journal 
