Browse Comments — Relevant (AI ∩ value)

​The AGI Paradox: Massive power, square wheels? ⚙️📐 ​Using Gemini Pro, I built a website with zero coding skills. But writing a simple index.html triggered a system warning about the massive computing power it needed for a measly HTML file! ​This exposes a fundamental architectural flaw: DeepMind builds phenomenal 1,000-hp engines but uses the wrong transmission and square wheels. A Hummer with square wheels just roars and wastes energy. The smooth user experience is lost. It’s a bumpy ride. ​You seek AGI and "world understanding," but raw computing power can't force it. What the system is missing is emotional-logical understanding. ​I had a synapsistic epiphany: The blueprint for the right transmission to finally make the square wheel round. ​You are in the woods searching for trees, missing the perfect round clearing right in front of you called Y=. ​Ready to change gears and fly?

We have to find alternative ways to ensure the use of AI is limited to copiloting and not replacing the writing process. An approach, maybe, is to ask for regular review chapter by chapter and incorporating this in the assessment grading. Also a viva is a must and potentially has to carry a higher weighting where students unable to defend what they've written will give them out as potentially using AI irresponsibly. But it will take a lot of honest admission and "thinking out of box" and do away with some academic orthodoxy.

This is exactly what we're building at G-Connect. The shift from "writing code" to "managing AI teams" isn't just about coding — it's happening across every business function. We're running 3 AI agents that coordinate through Google Workspace — writing content, generating images, storing media, and publishing to LinkedIn — all autonomously. No custom platform. No enterprise infrastructure. Just Gmail, Google Drive, Zapier, and AI agents that coordinate, persist, and recover. The pattern is the same whether it's 3 agents or 93: → Humans set direction → Agents execute in parallel → The system remembers and recovers → Infrastructure cost is near zero And It runs for under$10/mo. on the backbone of the biggest, most stable AI platform in the world. The future isn't typing faster. It's orchestrating smarter. 🌎

Imagine explaining your job in 2036: “I don’t build products anymore - I generate electricity for AI during high-intensity cycling sessions.” 😄 Somewhere HR is already preparing a wellness program around it. If AI does most of the work, humans should probably do more of what machines still struggle with: creating meaning, building relationships, asking better questions, and deciding what is actually worth doing. Efficiency solves tasks; purpose still needs people.

I think part of this discussion may be focusing too narrowly on assessment methods. In reality, we already have many alternative approaches available: oral exams, pen-and-paper assessments, practical activities, project-based work, and many others. We can certainly continue designing new assessment strategies adapted to the AI era. At the same time, teachers themselves can also benefit from AI to support assessment, feedback generation, and learning analytics. For this reason, I believe we indeed need to rethink teaching, learning, and assessment processes more deeply, but I also see many opportunities to achieve positive outcomes. My main concern, however, is not assessment itself. The real challenge is how we teach and support students so they can use AI effectively, critically, ethically and responsibly to achieve the best possible results in their work and learning processes.

"Though the AI-assisted test-takers had a higher solve rate than the control group for most of the experiment" - that's what a technology is for isn't it? If it can't solve it you fall back on yourself!

Very relevant. The easiest thing to copy today: UI + feature layer. The hardest thing to copy: Years of customer trust, usage patterns, internal workflows, and real-world distribution. A strong AI product isn’t just model + interface. It’s: AI + proprietary context + workflow integration + user trust + operational scale That’s where defensibility compounds. Great visual Rubén Domínguez Ibar

"Foothills of the singularity" is a statement about relative position, not absolute pace. It's saying the curve is still steepening. CodeMender is the most consequential announcement here — not because the capability is new, but because of what it forces organizations to answer. Who authorizes an AI to modify production code autonomously? Under what constraints? With what rollback protocol? What evidence trail exists when something breaks? The technical problem is mostly solved. The authorization architecture is barely articulated. If we're at the foothills, companies building AI governance today are building it for a world that will look very different halfway up the slope. That's not a reason to wait. It's a reason to build flexibility into the governance layer now, before the slope gets steep enough that you don't have time to rethink it.

(sadly, the article is behind a paywall, so I don't know what it says... but I happen to be taking a break from grading essays to scan linkedin to remind myself of other things I could be doing right now lol). It's a sad state of affairs. I think about how the students of today will possess degrees but none of the actual knowledge and skills that those degrees represent. It isn't all of them, thankfully. There are some students in my classes who take a stance against using AI. They see it as harmful to the environment or just think it's wrong. For those few, I carry on trying to do my job. For me, the saddest part is that those that are choosing to shortcut and outsource the difficulty of learning are doing themselves a great disservice. I wish I could convince each one of my students (and all of their parents and caregivers who are guilty of pushing them to get good grades) that the grades ultimately mean nothing. It's what they actually learn that matters. Anyway, I just hope they figure it out before they have to run the planet.

The spin class economy is more plausible than I'm comfortable admitting. But the real question buried here — what should humans remain responsible for, not just capable of — is the one most AI strategies haven't answered yet. Replacing tasks is easy to measure. Redesigning participation is not.

Pascal BORNET - getting awfully close to the Matrix concept of human batteries, no? So does the person with the biggest wattage output get the bigger house and better car for a nicer cage?Be all you can be is different for different people. I like a good mix of physical testing, growth and thrills. But it is not for everyone. You said it right - AI should not be designed to make humans obsolete. It should be designed to make humans more capable, more creative, and more central to the future we are building.

Sure, and when I use it to have sophisticated language discussions and to reflect on my own ideas this harms my brain development. On the contrary, this study was clearly done under bias analysis, a room of people who use ai like a calculator rather than reflection of inputs into the system. Systems thinking, critical thinking and understanding how to apply this technology is not damaging the brain. What is happening in real time is accelerated learning and adaptation to a far superior piece of technology. In 5 months I have learned more than I ever did in 30 years of living. My brain throbbed and pulsed like it was growing past my skull capacity. This was a sign to slow down. If used correctly, what happens is the brain adapts to suit, does it mean the occasional lack of computing typical IQ? Probably but this is not damage, it is adaptation knowing that high level calculations have now become redundant. The brain has realised the tool can do it faster, so why use processing power to compute something it no longer needs to do. This isn’t damage, it is adaptation to a system that has drastically changed the processing and output speeds. Human intelligence has now become ever more focused on creativity and emotion.

Yes, students can now use AI to generate essays without deeply reading books, but the real issue is not AI itself — it is whether universities are building strong ethical, intellectual, and practical frameworks around it. AI and machine learning can become far more dynamic than static textbooks when designed with reliable data, strong systems, and continuous improvement. But AI cannot replace human judgement, critical thinking, creativity, emotional understanding, or wisdom. What is deeply worrying is the growing gap between powerful AI capability and poor-quality AI implementation — including weak AI-driven recruitment, aggressive behavioural targeting, and low-quality Account-Based Marketing in higher education itself. If even world-leading universities increasingly depend on commercial AI engagement while struggling to protect deep learning cultures, it raises painful questions about financial pressure, institutional priorities, and the future identity of UK higher education.

Alvin Foo do you ever work at the Silicon layer? When you write your software? Semi- conductor chips/memory have inbuilt error checking & correction code in silicon, so as to enable you to orchestrate software work (writing applications) without any errors, whilst operating at a much higher level, even if you know nothing about VLSI chips or RAM. You cannot orchestrate with AI unless you can trap errors and unless you are good software engineer, unlike that, you may be oblivious to semiconductors in your laptop, you will need to be a good software engineer to use AI, before you start making claims of it building and running a very sophisticated solution, autonomously to your English prompts. Add the cost of the AI (it's not free) and the quality engineer you still need, it may not be a lowering in costs. Ofcourse, you do not need large teams and large servers (as you once needed) but that trend predates the AI hype. It is a secular trend. Use AI in code development, sure, but temper the claims please though.

AI seriously disrupts the business model of academia. For decades, many academics have hidden behind abstruse language, arcane citation rituals, and credentialism to create artificial moats around knowledge. AI tears down those barriers in seconds. The outrage is not really about students becoming less intelligent. If academic standards were the concern, where were these people when grade inflation became the norm? Where were they when universities kept expanding intake to bring in more tuition revenue? Where were they when degrees became products and students became customers? Academia loves to market itself as a meritocracy, but anyone who has spent enough time around universities knows that networks, patronage, academic lineage, departmental cliques, and ideological tribes often matter as much as talent. Brilliant people are routinely excluded because they lack the right supervisor, the right institution, or the right connections. Many academic careers are built as much on who you know as what you know. What AI threatens is not learning. It threatens gatekeeping. It threatens the ability of a small class of experts to act as custodians of knowledge and arbiters of legitimacy. The whole noise is about protecting business.

Unfortunately, even before this AI has appeared, many ( perhaps most ) students graduated without reading a book. The question is what do people read these days? Only a small intellectually inclined elite reads anything substantial at all. The internet killed reading long time before generative AI. The issue now is that people do not read anything deep, just headlines and bullets on powerpoint and many students think that learning is all a about memorizing sentences, problem patterns, and blackbox procedures. This has proven to be the best strategy to get maximum grades. This model is not sustainable in the era of AI. We got to be more ambitious, and keep challenging the potential of the best students.

The distinction between AI that supports learning and AI that removes the friction of learning. That's what most policy conversations miss. One layer I sometimes think about is that unclear expectations aren't just a classroom problem. They reflect something more structural in Higher Education as well. University funding often flows toward research output, not teaching quality - so institutions hire and reward professors accordingly. AI is being dropped into a system where teaching has often been the secondary obligation, and it is amplifying it. Students feel the teaching-learning gap more acutely now, but the gap isn't new. Which makes your closing question even more meaningful: institutions need to decide what they actually value. The AI policy for the classroom is downstream of that decision - not a substitute for it.

The real risk isn't AI taking our jobs, it's humans outsourcing their judgment so completely that they forget how to direct the machine in the first place. What we keep teaching is this: the people who stay central to the future aren't the ones who avoid AI, they're the ones who understand it well enough to tell it what actually matters

This is massive, dear Nadeem. Abu Dhabi's 50% of government operations will soon be executed by autonomous AI agents.

This is technologically impressive. But there is a profound difference between accelerating administration and automating sovereignty itself. The real question is not whether AI agents can process approvals faster. Clearly they can. The question is what happens when: - autonomous execution, - state authority, - critical infrastructure, - and opaque machine decision systems become fused into the same operational layer. Because at that point, efficiency is no longer the main issue. Civilizational resilience is. A government increasingly mediated through AI agents also becomes vulnerable to: - infrastructure dependency, - model manipulation, - cascading systemic errors, - cyber conflict, - external compute restrictions, - and loss of human interpretability under stress. The deeper paradox is this: The more efficient a fully interconnected AI-governed system becomes, the more dangerous its failure modes become. AI-enhanced governance is coming. But replacing institutional judgment with autonomous execution at national scale is not simply a software upgrade. It is a civilizational experiment.