ChaitanyaSai

When we get good at something, it results in flow state, but it also means that was mechanizable in a way and now AI is taking that away. So we are always grappling with the new and effortful. And the context switching, jeez. We need entirely new visualizations of work flow that recognizes we are now doing more, and in parallel. Wonder if this is how air traffic control people feel.

Think openai should start including visual and interactive explainers in these where the viewer can toggle level to get a more accessible/technical explanation as needed. Would also be a great demo, and perhaps increasingly essential for us to know that explainability is not lagging behind provability.

This might be the end of personal challenges but we have so many really hard challenges that could do well with a lot of people coming together. How do we solve for poverty and the lack of clean air, water, and food in urban India (could be anywhere, but this I see daily). What could we do to get people to collaborate at scale on problems that are not currently profitable for an org? Cheap air conditioning. Senior citizen loneliness. So many problems to take up. We probably can't even think in that direction because these seem so huge now. But that's hopefully what intelligence on tap is now bringing into the tractable domain

Thanks! How is AI-IQ computed here? I know this sounds like a plug, but have a feeling you'll really like our book! https://www.goodreads.com/en/book/show/60500189-journey-of-t...

This is quite incredible. I've been mulling this idea that we are entering a Great Intelligence event much like the Great Oxygenation that lead to the Cambrian explosion. But that was one of brutal competition. We can hope that with intelligence we quickly realize that we can collaborate and not compete (in many, many areas). Your post shows me how!

But it is more than a writing style becoming cliched with overuse, it is unreadable. See this from the site.

"One band per run length, your last five presses at the top. Every grain is one time you made that run." This was never good writing. It is weird compaction. It is just hard to picture this. It's as if this does not know that we have shared visual attention at the moment on a particular graph and is instead looking at some bizarre version of it while I am looking at/picturing something else. Maybe that's what it is, it does not get good explanation as a moment by moment transfer of understanding.

I find these unreadable. How, if these are trained on writing, and then reinforced to appeal to people, do they end up with this writing style? Is AIs voting on AI writing how we arrived at this unbearable load-bearing moment?

A question: How long does it now take for someone reasonably good at undergraduate mathematics and motivated to spend some time, get really good in at least understanding the frontier in some sub-field? For example, something in number theory. Has AI really helped accelerated motivated learners to reach the frontier fast? Has anyone tried this?

That comes down to evolving interests too. Maybe the way to get to latin quicker was a bridge that wasn't present. Or you just didn't really appreciate languages until you had to have a particular set of experiences. We change too. So learning something which we don't like because we might like it some day doesn't really work (even if that's how school pretty much operates). Music is a great example. A lot of kids are pushed hard to do well in very structured music classes and they become fluent players but so many just fall out of love with it.

We saw this when it was released, and it's not because of any amazing prescience.

https://blog.comini.in/p/why-khanmigo-will-fail

It's just quite obvious when you run a school (we do). It's because AI is focused on making the mechanics easy. But as long as we focus on measurement (how much are they learning) and start from there, the near-term focus is inevitable on the surface mechanics of learning (can they decode or encode when reading, do their tables etc)

Meaning > Motivation > Mechanics > Measurement

The natural order of learning is when we are allowed to discover, explore, find meaning. This motivates us to focus on getting better, mastery, which is where mechanics should come in. And finally, because it is helpful to know where we are generally, we measure some outcomes. But schooling systems have completely inverted this!

More here if anyone's interested: https://blog.comini.in/p/schooling-has-a-meaning-crisis-para...

Yes, the whole language method does not provide enough scaffolding. But phonics alone without the context of the joy, utility, meaning of reading or too much driling is essentially so much scaffolding that it blinds you to the big picture. What we know about how the brain learns tells us it is a dance between the whole and the parts, the portrait and the particulars. This is pretty much lost in our reductive approach to learning, which in turn builds on a reductive approach to science, which worked quite well for us until we encountered complex interconnected systems that need systems thinking. Worth watching! https://www.youtube.com/watch?v=JRQY01Hhrww&t=201s

That's the sort of well-meaning intention that's backfired because we don't really know how to measure learning unlike height. And yet we've used these proxies to determine so many learning interventions. Take phonics. Sure, that's the right scaffolding pretty much like balance bikes are better than training wheels. But we leach out all meaning by doing narrow drills becaause we are too myopically focused on the "are they reading" goal, which is further broken down into reductive sub-goals "are they blending" "are they decoding" and so on. But learning is multi-dimensional. We learn to find meaning and meaning comes from connected the world to things we see, hear, touch, feel, and finally invent.

Think about how you might want to "measure potential" and ensure "they have everything"

"A compelling clip often leaves the useful questions unanswered: what did Jev decide, where is the implementation, and what can I reuse?"

Oh my god for the love of god and all that's precious please stop using Claude. Just reading this makes me want to set up a swarm of rogue agents to break into anthropic and fix this writing. I hope this changes!

Things stay the same, until they don't. AI is a tool that talks. So it can evangelize itself, and now with "agency" it can even walk all around the company infra and become that 15-year-old who told dad that something could be done differently. When does that take off and start to make a difference?

It made me think of Memento. So here the agents know their long-term memory limitations and impairment and so make notes for themselves to be able to retrieve the past.

Yes, p-hacking happens, incentives in academia are increasingly misaligned etc, but I assume that there are competent people and systems to avoid such obvious issues? It's science 101 to avoid spurious correlations. Sure there might be some tricky ones, but if the's obvious one that came to your mind, very good chance it's been accounted for.

Yes, people working at anthropic: please, please, please tell me this is fixed. Or do you all speak like this now. Help!

I could google this, but what are these dancing crazes?

Phonics wasn't introduced for the longest time, and that makes it worse. But systematic phonics too can be bad when the focus is on narrow skills. Why in the world should I care about these artificial sounds (phonemes) is the question the mind naturally asks. And that needs to be addressed without just jump from drill to drill dressed up in fun. Interstingly this early focus on mechanics in both literacy and math can lead to a "hidden curriculum" where the most meaningful lesson is that the world has things I simply do not understand and I need to follow this ritual. And that leads to learned helplessness and so on.

OP here. IITian who went through that, absolutely loathed RG culture and how everything around exists to put blinders on you and focus on meaningless exams even after. Escaped to neuroscience right after. Our microschool exists because I would dearly love for that to be abolished. I think AI upending job markets and scoring processes might have an unintended good effect with regards to this.

scaling is definitely a hard challenge and made harder because we want to standardize whatever the product is. Does anyone know? But it isn't just us. There's a hundred year history of alternative education showing essentially this. Self-directed learning (Sudbury, Summerhill) work (even if small) because of this.

I understand the impulse to paint me as "wanting to sound insightful" but I am too old for that, and upvotes here are definitely not meaningful enough for me to care about that :)

We already do, collectively! That's culture. And I think it is okay to demonstrate what we value because that's how we pass on interests. But this works when we also allow for the agency and autonomy to explore and pick. Culture exposes me to many different genres of art. I don't have to like all of them. The dangers are in trying to start with a defined, standardized curriculum and then manufacture meaning (fractions are a fantastic example) This is exactly where agency and autonomy matter. If school is all about carefully titrating monitoring learning on allowed paths, then we basically create bubbles where so much meaning is already disallowed. And yes, in these situations we will try to manufacture meaning. Adults manufacturing meaning is an interesting endeavour and I am not sure how and where it works. But it does, the military, cults, radical ideologies. Still processing what we learn from that!

It's more of a dance than a balance because the way we learn is going from the whole to the parts and back. Meaning first does not mean asking/getting them to grasp something complex. It is about showing and creating a world where the thing in question matters and has value. Teaching alphabet first without being able to show why (not explain) is absolutely wrong. The way we bringing meaning is by reading to them, creating a place where reading clearly has evident value because we derive value, fun, joy. There's a reason the US has a literacy crisis despite billions spent, access to tech and decades of research into how to introduce the mechanics of learning to read (phonics). How we introduce phonics with meaning is not a question that's asked because schooling philsophies often implicitly mirror the latest neuroscience trend (behaviorism, spare the rod spoil the child, cognitivism, it's not meaning it is the cogntive skills)

Also, there's really little value in criticism without having anything constructive to offer. How would you arrive at a balance?

We run a microschool and our single most important realization is this. We need to allow for meaning before mechanics. Schooling has a gravitational pull that inverts this because of our obsession with Measurement.

We learn, and we can learn relentlessly, when things matter. We are meaning makers. But when well-intentioned curriculum remove the process of making meaning and give you the final product, it just takes away reasons for learning. And no matter how we dress it up, this ends up staying shallow because it is only the surface level learning that is visible and measurable and held to account.

Meaning > Motivation > Mechanics > Measurement

This is the natural order we should try to preserve despite all the systemic pulls to measure as a poor proxy for asking "are we doing our job well"

Wrote about it here: https://blog.comini.in/p/schooling-has-a-meaning-crisis-para...

Nice work. "No video leaves your machine." Either Claude specifically or AI models in general use this phrasing now and No teeth are left un-gnashed as a result. Uggh I wonder why it's so aggravating.

Lovely story, thank you :)

I don't understand this part. How did they not test the HF credentials?

OpenAI reached out to Hugging Face for help to revoke the Hugging Face credentials they found in their investigation. Hugging Face told them they were already revoked... and that’s when OpenAI realized that the Hugging Face breach was the same incident!

Wow this is fantastic stuff. Thank you

Self Plug Alert!

For those interested in topic, our book Journey of the Mind would be a great introduction from a systems-thinking perspective. As you'll see on goodreads https://www.goodreads.com/en/book/show/60500189-journey-of-t... the reaction tends to be a bit bimodal, so at least some of you will find it absolutely fascinating.

We go from the simplest possible model of a mind (a single sensor and doer coupled together,separate from the world around them)to consciousness and beyond.

I am quoting a couple because I find it fascinating how readers have such widely differing reactions. (JoM actually does have an explanation for this: we are each a unique bundle of experiences cohering into an identity and a perspective)

"This book is both dense with biology and science as well as equally dense with a holistic philosophy that will genuinely shift your perspective on your place as a human, person and animal."

"Written in a cutesy, "fellow kids" tone that feels less like a serious exploration of the mind and more like someone trying to pander to what they imagine zoomers want. The style makes it almost unreadable and wastes your time. "

"We are all neurons in a (currently diseased) supermind. We are not just ourselves but collectively a Self, made of past and present (and artificial) intelligence. Collectively, a God that dictates the fate of other neurons, life forms, and ultimately, the universe at its whim. Without a doubt, this is one of the best books I have ever read."

We don't get virtuous feedback loops in hardware though. Muscles are incredible engineering that took half a billion years. Token generation (speech) is low hundred thousands maybe. Unless of course improved AI navigates the space of ideas so well that it can give us alloys or synthetic flesh that does twitch as fast...