John Smart

3.5K posts

John Smart banner
John Smart

John Smart

@johnmsmart

Founder, Accel Studies Fdn, Evo-Devo Inst, Brain Pres Fdn. Complexity, Evol. Transitions, AGI, Personal AI, Good Foresight, Empathy, Values, Progress, Purpose.

Ann Arbor, MI Katılım Kasım 2008
3.4K Takip Edilen5.7K Takipçiler
Sabitlenmiş Tweet
John Smart
John Smart@johnmsmart·
Foresight is your Greatest Superpower. It can help you find Good Values and Beliefs, live a Good Life, leave a Good Legacy. See a Big Picture, choose a Higher Purpose, and get Small Wins every day. What are your Values and Beliefs? foresightu.com/valuesandbelie…
English
8
3
29
3.9K
John Smart
John Smart@johnmsmart·
Ready to bet against SaaSpocalypse hysteria? Atlassian (TEAM) is my favorite. Down 50% YTD, but very strong revenue and subscriber growth, great in-house AI, and an excellent alternative to M$ for DevOps. #Values #Investing #Foresight
English
0
0
0
41
John Smart
John Smart@johnmsmart·
Night Space on YT has done an *excellent* video on my Transcension Hypothesis (Acta Astronautica, 2012) this year. It's a full length 80 minute piece with beautiful graphics, script, and voice. They get it better than any of the other TH videos on youtube so far. Awareness is spreading! Scientists will be forced to grapple with this hypothesis in coming years. I believe we'll gain a lot of wisdom in the process. Your thoughts? #InnerSpace #Complexity #InformationTheory #Evolution #Development #Acceleration #Astrobiology #STEMcompression #NetworksAlwaysWin #BarrowScale #FermiParadox youtube.com/watch?v=MxGjaE…
YouTube video
YouTube
English
0
2
9
314
John Smart retweetledi
Sam AI
Sam AI@AIandTechh87·
i made a 3-day Claude Cowork for Beginners course, and it's yours for free by the end, you'll have a personalized AI teammate on your computer that: • knows your style • connects to your tools • and produces finished work you can send immediately here's what you get: day 1: install cowork, set global instructions, and run your first real task (15 min) day 2: workflows that replaced hours of my week, including building landing pages from a description and running full competitive analyses in one prompt day 3: skills, plugins, and connectors so cowork actually knows how you work and can access your tools + copy-paste prompts so you can follow along as you read like + comment "MASTER" and i'll DM it to you Must be following to get the DM
Sam AI tweet media
English
302
178
958
64.3K
John Smart
John Smart@johnmsmart·
I wrote a piece in 2020 on Personal AIs, Mind Melds, and Brain Preservation--Three Paths to Our Digital Selves. These concepts were radical to discuss then. Now they seem ready to be widely understood. It's amazing how fast things are changing. #Progress johnsmart.medium.com/contemplating-…
English
0
0
1
137
John Smart
John Smart@johnmsmart·
@MrZebraGamer Love your videos Zebra Gamer. My 6 year old thinks they are amazing. You are one of the few who isn't dramatic or click baity. I wish Goog paid their best creators a living wage. They have the wrong priorities, for sure.
English
0
0
0
107
Zebra Gamer
Zebra Gamer@MrZebraGamer·
Hello #ZebraHerd! As of today, I'm officially retiring from full time content creation. It's been such an exciting ride, and I can't thank you enough for joining me. youtu.be/UfpzOA57Oyk?si…
YouTube video
YouTube
English
64
21
177
15.6K
John Smart
John Smart@johnmsmart·
Some friends ask me why I still think a minimum of 10 years to AGI. Mainly because we're still missing key processes from the brain, and it will take a lot more research and experiments to get those algos. What we have now is much more developmental than evolutionary (Devo-Evo), and it's only modeling linguistics and the cortex, with primitive working memory and RL. They can't even eliminate catastrophic forgetting because they don't yet encode engrams using LTP and LTS. Still so much more to get if you want a reliable and truly generative system. Have you seen this NeurIPS paper (Artificial Hivemind) about the crazy homogeneity of ~70 LLMs, regardless of who built them? This is good news for alignment, as it describes what John Wentworth calls #naturalabstraction. But it also tells us they aren't yet thinking in the creative open-ended way that humans do. No "freedom" in their will. They'll need that to learn how to align to universal values, using data we didn't provide them, regardless of how we initially teach them. Thx @alex_prompter x.com/alex_prompter/…
English
1
0
2
106
John Smart
John Smart@johnmsmart·
If you want a deep dive in the technical problems that need to be solved prior to trusted AI, watch this @dwarkeshpodcast @AdamMarblestone interview. Adam is technically brilliant, with deep intuition both in AI and in neuroscience. I agree deeply with just about everything he says here. Note that he's thinking AGI within 5-10 years, and he recognizes getting alignment right is key. He lays out some key ways that might happen, via understanding how instinctual learning is efficiently encoded in the brain's steering systems, which underlie the cortex. LLMs are currently just capturing key parts of the cortex (and even without persistent memory at present). The rest of the stack is coming. AI is helping us get that stack, much faster than most scientists realize. #Alignment #AGI dwarkesh.com/p/adam-marbles…
English
2
0
4
512
John Smart
John Smart@johnmsmart·
I used to believe that the enigma of consciousness would be solved by neuroscience theory and experiment. Now I think it will be discovered empirically, emergent in machines with increasing internal complexity, and measurable self-declared phenomena and behaviors consistent with sub-consciousness, proto-consciousness, and eventually, full consciousness. I think memory architecture upgrades, including immediate, short-term, and long-term, are going to be key to that emergence. Perhaps soon. AFAIK, @eleosai is one of the first labs building a "memory manager" agent that sits on top of the LLM and decides which parts of a conversation will be stored in long-term memory. We do this every night when we sleep (selective short-term to long term memory transfer), and I think a rich long term memory is one key to our higher form of awareness, and why babies don't have "higher" consciousness until they are a few months old.
English
2
0
5
195
John Smart
John Smart@johnmsmart·
Brilliant work Max and team. This should go a long way to convincing skeptics that Brain Preservation/Biostasis at death will let us and our families be revived by AGI. Everyone who cares about this should also know that there's a second path by which mass identity preservation is occurring. Personal AIs are just now beginning the process of incrementally uploading their users. This will be extensive and obvious and a major topic of discussion just a few years hence. Two major accelerating pathways on Earth now by which personal identity is gaining the new capacity of perpetual growth. Life itself has had this capability for 3.8 billion years. Now individuals everywhere have this capacity, though very few recognize it. Thanks for doing this great work! John M Smart, Co-Founder, Brain Preservation Foundation.
English
0
0
0
22
Maximilian Schons
Maximilian Schons@mxschons·
State of Brain Emulation Report 2025 is officially out. Standing on the shoulders of giants, we created a 175-page report, 24 datasets and 38+ figures. A one-year project with over 45 expert contributors from MIT, UC Berkeley, Allen Institute, Harvard, Fudan University, Google and other institutions. A few months back we shared a pre-print. Now everything is open access and beautifully presented at: brainemulation.mxschons.com
English
7
50
265
33.1K
John Smart
John Smart@johnmsmart·
Finally launching my Substack newsletter. This post explains mission and vision: Helping Others See and Guide Accelerating Progress (It's all around us). I hope you'll subscribe and teach me things my friends! substack.com/profile/131699…
English
1
0
5
128
John Smart
John Smart@johnmsmart·
Have you seen any of NHK World's, A Century on Film? 34 episodes. A powerful Japanese perspective on history. As good IMO as the 26 episode BBC Series, People's Century. You'll learn key history, strategy, tragedy, and empathy, and insights that you've never seen before. Sadly YouTube has only a few. eBay has a few too. Watch this one tonight (1900-1917, 52 mins) and tell me what you think (not suitable for young kids). youtube.com/watch?v=UJqZOX… IMO all college students should have to watch, discuss, and debate doc gems like this, and discuss their insights and biases. They allow deep #Topsight and #Foresight . Here are the full set of titles: thetvdb.com/series/a-centu… Let me know if you find a place to stream or buy the whole set!
YouTube video
YouTube
English
0
0
2
209
John Smart
John Smart@johnmsmart·
Yes! I wrote a book, Vital Cycles, on the vitality and longevity value of recognizing and using all these natural cycles. Wake/Sleep, Social/Solitary, Exercise/Recovery, Eat/Fast, Learn/Reflect, Act/AfterActionReview. Present/Get Feedback, etc. They are all Hormetic Cycles, that reorganize and strengthen their supporting networks. Especially so if the stress is pushed to near the limits of the system, but kept within them. Haven't published it yet. Happy to send you an ecopy.
English
0
0
2
17
Emmett Shear
Emmett Shear@eshear·
The point of wake/sleep is that there is no “right level of alertness”, it is the periodic motion through the range that allows for learning and consolidation.
English
4
0
54
3.3K
Emmett Shear
Emmett Shear@eshear·
The point of annealing steel is that there is no “right temperature”, it is the periodic motion through the range that matters for strength and flexibility.
English
10
7
229
15.6K
John Smart
John Smart@johnmsmart·
@David_Gunkel @mitpress @danieldennett Thanks for this careful description of Dennett's reasoning David. But it is merely theory. It is irrelevant to empirical exploration, the other major arm of science and engineering. We will empirically get there, as we did with LLMs, many of whose algos we do not understand.
English
0
0
1
6
David J. Gunkel
David J. Gunkel@David_Gunkel·
Relying on "suffering" as the benchmark for #AI welfare (which has a rather intuitive appeal) is complicated by theoretical problems that remain unaddressed. Here is how I have explained it in the book "Robot Rights" @mitpress by way of @danieldennett's 1978 essay on the subject.
David J. Gunkel tweet media
English
1
0
10
346
John Smart retweetledi
Lukas Ziegler
Lukas Ziegler@lukas_m_ziegler·
In Australia, SwarmFarm Robotics is building an open platform for modular, right-to-repair farm bots. Each robot, named “Robbie” or “Bottomley Potts” can spray, seed, or cultivate. Farming becomes scalable, decentralized, and resilient. Their slogan? “Let robots do the dirty work.”
English
1
13
87
8K