playing with Perplexity (a chatbot); sounding it out about ISEPP themes
24 views
Skip to first unread message
kirby urner
Feb 15, 2025, 8:11:25 PMFeb 15
to wwwan...@googlegroups.com
Hey guys, I know if we were meeting weekly we'd be yakking about AI quite a bit, given we're such nerds 'n all.
Well, along those lines, I am here to heap praise on Perplexity, a front end to some chatbots, I'll do some research about which ones if people are interested.
I know they added DeepSeek R1 to their Pro search, of which I get about three each month. The one below was a plain vanilla freebie to me.
Visit perplexity.ai if you wanna know more, or click my link at the bottom (which also takes you there).
You'll find my prompt resonates with ISEPP themes, including featured books and speakers.
I've stumbled on a new-to-me economist, Keen, and I feel inspired by a lot of what he says just as Terry was so inspired by Romer.
Then from my Quaker background, I knew about Kenneth Boulding already. Might we tie them together, along with that Dorion Sagan lecture?
I'm impressed by how good the chatbot's reply is.
I'm not doing any fixing up whatsoever other than to make the formatting prettier (e.g. real boldface instead of some markup punctuation).
Terry, are you listening? This reply is great, no?
---
Prompt (written by me):
How do the thermodynamically aware views of Steve Keen dovetail with (a) those of general systems theorist Kenneth Boulding and (b) those expressed in the book Into the Cool, which talks about our energy economy and ecosystem riding a solar gradient. The planet is already fusion powered in that sense.
Reply (what I got back in a few seconds):
Steve Keen's thermodynamically aware views on economics align closely with both Kenneth Boulding's general systems theory and the concepts presented in "Into the Cool" by Eric D. Schneider and Dorion Sagan. These perspectives share a common thread of incorporating energy and thermodynamic principles into the understanding of complex systems, including economic and ecological processes.
Alignment with Kenneth Boulding's General Systems Theory
Keen's approach to economics resonates with Boulding's hierarchy of system complexity:
1. Keen emphasizes the role of energy in economic models, which aligns with Boulding's recognition of open-throughput systems[2]. Both acknowledge that economic systems involve continuous flows of energy and materials.
2. Boulding's hierarchy progresses from simple mechanical systems to complex biological and social systems[5]. Similarly, Keen argues for moving beyond simplistic neoclassical economic models to more complex, dynamic representations that account for energy flows[1].
3. Both thinkers recognize the limitations of reductionist approaches when dealing with higher levels of complexity[5]. Keen's critique of neoclassical economics for ignoring energy aligns with Boulding's warning against applying lower-level concepts to more complex systems.
Connections with Into the Cool
Keen's work dovetails with the concepts presented in "Into the Cool" in several ways:
1. Energy as a fundamental driver: Both Keen and the authors of "Into the Cool" emphasize the critical role of energy in driving complex systems, including economies and ecosystems[3][6].
2. Thermodynamic principles in economics: Keen's incorporation of energy into economic models echoes the book's application of the second law of thermodynamics to various fields, including economics[3].
3. Complexity and energy gradients: "Into the Cool" discusses how complex systems emerge to dissipate energy gradients[6]. This aligns with Keen's view that economic models should account for energy flows and their role in economic processes[1].
4. Interdisciplinary approach: Both Keen and the authors of "Into the Cool" advocate for bridging disciplines, particularly connecting physics (thermodynamics) with economics and ecology[3][7].
The concept of the planet being "fusion powered" through the solar gradient is consistent with Keen's emphasis on energy's role in economic systems. This perspective recognizes that the Earth's ecosystems and human economies ultimately depend on the energy flow from the sun, which is indeed a product of fusion[9]. Keen's work, by incorporating energy into economic models, implicitly acknowledges this fundamental energy source driving our planet's systems.
In conclusion, Steve Keen's thermodynamically aware economic views align well with both Boulding's systems theory and the concepts in "Into the Cool." All three perspectives emphasize the importance of energy flows, complexity, and interdisciplinary approaches in understanding economic and ecological systems, moving beyond traditional reductionist models to more holistic, energy-aware frameworks.
Citations:
[1] https://www.rebuildingmacroeconomics.ac.uk/post/integrating-macroeconomics-and-ecology-via-energy-and-the-laws-of-thermodynamics
[2] https://www.beyondintractability.org/moos/system-levels
[3] https://press.uchicago.edu/ucp/books/book/chicago/I/bo3533936.html
[4] https://exergyenergy.com/financial-thermodynamics/
[5] https://systemsthinkingalliance.org/understanding-the-boulding-hierarchy-of-complexity/
[6] https://archive.org/details/intocoolenergyfl0000schn
[7] https://profstevekeen.substack.com/p/the-role-of-energy-in-production
[8] http://grahamberrisford.com/AM%204%20System%20theory/SystemTheory/ChallengingSystemsThinkers/01%20Bouldings%20ideas.htm
[9] https://press.uchicago.edu/Misc/Chicago/739368.html
Alignment with Kenneth Boulding's General Systems Theory
Keen's approach to economics resonates with Boulding's hierarchy of system complexity:
1. Keen emphasizes the role of energy in economic models, which aligns with Boulding's recognition of open-throughput systems[2]. Both acknowledge that economic systems involve continuous flows of energy and materials.
2. Boulding's hierarchy progresses from simple mechanical systems to complex biological and social systems[5]. Similarly, Keen argues for moving beyond simplistic neoclassical economic models to more complex, dynamic representations that account for energy flows[1].
3. Both thinkers recognize the limitations of reductionist approaches when dealing with higher levels of complexity[5]. Keen's critique of neoclassical economics for ignoring energy aligns with Boulding's warning against applying lower-level concepts to more complex systems.
Connections with Into the Cool
Keen's work dovetails with the concepts presented in "Into the Cool" in several ways:
1. Energy as a fundamental driver: Both Keen and the authors of "Into the Cool" emphasize the critical role of energy in driving complex systems, including economies and ecosystems[3][6].
2. Thermodynamic principles in economics: Keen's incorporation of energy into economic models echoes the book's application of the second law of thermodynamics to various fields, including economics[3].
3. Complexity and energy gradients: "Into the Cool" discusses how complex systems emerge to dissipate energy gradients[6]. This aligns with Keen's view that economic models should account for energy flows and their role in economic processes[1].
4. Interdisciplinary approach: Both Keen and the authors of "Into the Cool" advocate for bridging disciplines, particularly connecting physics (thermodynamics) with economics and ecology[3][7].
The concept of the planet being "fusion powered" through the solar gradient is consistent with Keen's emphasis on energy's role in economic systems. This perspective recognizes that the Earth's ecosystems and human economies ultimately depend on the energy flow from the sun, which is indeed a product of fusion[9]. Keen's work, by incorporating energy into economic models, implicitly acknowledges this fundamental energy source driving our planet's systems.
In conclusion, Steve Keen's thermodynamically aware economic views align well with both Boulding's systems theory and the concepts in "Into the Cool." All three perspectives emphasize the importance of energy flows, complexity, and interdisciplinary approaches in understanding economic and ecological systems, moving beyond traditional reductionist models to more holistic, energy-aware frameworks.
Citations:
[1] https://www.rebuildingmacroeconomics.ac.uk/post/integrating-macroeconomics-and-ecology-via-energy-and-the-laws-of-thermodynamics
[2] https://www.beyondintractability.org/moos/system-levels
[3] https://press.uchicago.edu/ucp/books/book/chicago/I/bo3533936.html
[4] https://exergyenergy.com/financial-thermodynamics/
[5] https://systemsthinkingalliance.org/understanding-the-boulding-hierarchy-of-complexity/
[6] https://archive.org/details/intocoolenergyfl0000schn
[7] https://profstevekeen.substack.com/p/the-role-of-energy-in-production
[8] http://grahamberrisford.com/AM%204%20System%20theory/SystemTheory/ChallengingSystemsThinkers/01%20Bouldings%20ideas.htm
[9] https://press.uchicago.edu/Misc/Chicago/739368.html
---
If you got this far, then wait, there's more:
That should be a link to the whole thread, where I start out prompting it to cough up Steve's name but it coughs up Eric's instead... anyway, you'll see.
Kirby
kirby urner
Feb 15, 2025, 10:02:49 PMFeb 15
to WWWanderers
Just to quickly follow up in case you're curious, I asked Perplexity itself to explain what large language models it's consulting. I don't think it's answer is necessarily up to date in terms of Deep Seek R1 which I earlier saw had been added as a part of the Pro lineup.
Here's that query, you shouldn't have to sign up or give your email (if a box pops up, just close it):
Prompt:
How is Perplexity powered in terms of which LLMs is it a front end for?
Answer:
Perplexity AI utilizes a combination of proprietary and third-party large language models (LLMs) to power its search and answer capabilities. The specific LLMs used by Perplexity include:
1. Proprietary models:
- PPLX-7B-Online: A 7 billion parameter model based on Mistral 7B[7].
- PPLX-70B-Online: A 70 billion parameter model based on LLaMA 2 70B[7].
- Custom fine-tuned models optimized for Retrieval Augmented Generation (RAG)[6].
2. Third-party models (available in Perplexity Pro):
- GPT-4 Omni: OpenAI's latest model, known for its strong reasoning abilities[9].
- Claude 3.5 Sonnet and Haiku: Anthropic's advanced LLMs, with Sonnet excelling in nuanced language tasks[9].
- Grok-2: Mentioned as available in Perplexity Pro[3].
- Llama 3: Also listed as accessible through Perplexity Pro[3].
Perplexity's free version primarily uses their standalone LLM based on GPT-3.5 with web browsing capabilities[3]. The Pro version provides access to a wider range of models, including those from OpenAI, Anthropic, and other providers, allowing users to leverage different AI capabilities for various tasks[3][9].
Citations:
[1] https://blog.zingacp.com/2024/05/20/what-llm-does-perplexity-ai-use/
[2] https://dev.to/paka/i-built-perplexity-ai-with-nextjs-and-open-source-llms-1gl3
[3] https://en.wikipedia.org/wiki/Perplexity_AI
[4] https://blog.phospho.ai/how-does-ai-powered-search-work-explaining-rag-llm-and-perplexity/
[5] https://bdtechtalks.substack.com/p/will-this-be-perplexity-ais-moat
[6] https://www.reddit.com/r/LocalLLaMA/comments/1bh6o3e/reverse_engineering_perplexity/
[7] https://www.perplexity.ai/hub/blog/introducing-pplx-online-llms
[8] https://www.forbes.com/sites/joannechen/2023/09/06/how-perplexityai-is-pioneering-the-future-of-search/
[9] https://www.perplexity.ai/hub/technical-faq/what-advanced-ai-models-does-perplexity-pro-unlock
1. Proprietary models:
- PPLX-7B-Online: A 7 billion parameter model based on Mistral 7B[7].
- PPLX-70B-Online: A 70 billion parameter model based on LLaMA 2 70B[7].
- Custom fine-tuned models optimized for Retrieval Augmented Generation (RAG)[6].
2. Third-party models (available in Perplexity Pro):
- GPT-4 Omni: OpenAI's latest model, known for its strong reasoning abilities[9].
- Claude 3.5 Sonnet and Haiku: Anthropic's advanced LLMs, with Sonnet excelling in nuanced language tasks[9].
- Grok-2: Mentioned as available in Perplexity Pro[3].
- Llama 3: Also listed as accessible through Perplexity Pro[3].
Perplexity's free version primarily uses their standalone LLM based on GPT-3.5 with web browsing capabilities[3]. The Pro version provides access to a wider range of models, including those from OpenAI, Anthropic, and other providers, allowing users to leverage different AI capabilities for various tasks[3][9].
Citations:
[1] https://blog.zingacp.com/2024/05/20/what-llm-does-perplexity-ai-use/
[2] https://dev.to/paka/i-built-perplexity-ai-with-nextjs-and-open-source-llms-1gl3
[3] https://en.wikipedia.org/wiki/Perplexity_AI
[4] https://blog.phospho.ai/how-does-ai-powered-search-work-explaining-rag-llm-and-perplexity/
[5] https://bdtechtalks.substack.com/p/will-this-be-perplexity-ais-moat
[6] https://www.reddit.com/r/LocalLLaMA/comments/1bh6o3e/reverse_engineering_perplexity/
[7] https://www.perplexity.ai/hub/blog/introducing-pplx-online-llms
[8] https://www.forbes.com/sites/joannechen/2023/09/06/how-perplexityai-is-pioneering-the-future-of-search/
[9] https://www.perplexity.ai/hub/technical-faq/what-advanced-ai-models-does-perplexity-pro-unlock
kirby urner
Feb 17, 2025, 7:49:37 PMFeb 17
to Terry Bristol, wwwan...@googlegroups.com, Terry Bristol
Hi Terry --
I'm tempted to track down Keen's email and zap him the Perplexity query, as a way of introducing where I'm coming from.
Better yet: a link to this very thread on our world-readable Google listserv (but only members can post) would be in order.
I watched a YouTube about Perplexity's new "deep research" option, available on free tier (i.e. my tier) and thought I'd give it a try re the prompt:
---
Prompt:
In his World Game workshops, Buckminster Fuller proposed humanity keep working on a global grid, connecting its local electrical grids into one big global network. A nonprofit called GENI took up the cause, spreading awareness of the proposal. Some Chinese engineering companies took a favorable attitude. Yet there has been political pushback and doubts have been raised about how realistic this proposal might be. Fuller cast his proposal as a part of a “design science revolution” wherein humanity would take care of its physical needs in a successful manner. What is the current state of this proposal and who are its primary advocates today?
Deep Research response by Perplexity:
---
[that's from https://groups.io/g/synergeo/message/2955 -- another publicly archive listserv on which I'm active ]
Kirby
On Mon, Feb 17, 2025 at 8:08 AM Terry Bristol <terr...@gmail.com> wrote:
Kirby –OK. You have my attention. (Just what I needed – another exciting distraction.)I will want to look at Keen in particular.Perplexity seems to gather a lot of ‘relevant’ information, but I don’t see any ‘intelligent’ addition.Are you proposing an in-person or online dialogue on all this?I am up for both.Let me knowTerry
kirby urner
Feb 18, 2025, 2:37:07 PMFeb 18
to Terry Bristol, Wwwan...@googlegroups.com
Thanks for breakfast! I learned a lot. Enough to make up a query (below).
Check it out:
(a Perplexity Deep Research computation)
Prompt:
Adrian Bejan has interesting insights into thermodynamics. Do people ever make a connection between his writings and the economic theories of thermodynamically aware economists such as Steve Keen? How would we connect Bejan’s thinking to that of Paul Romer? What other economists should we be thinking about in connection with Romer and Bejan?
I found the chatbot's report eerily (uncannily) on the money, albeit my coming from an intelligent layman's perspective, not an economist's.
Kirby
On Mon, Feb 17, 2025 at 9:22 PM kirby urner <kirby...@gmail.com> wrote:
FYI you'd likely find this ISEPP-like lecture somewhat interesting:It's my latest YouTube wherein I talk through a deck of Google slides showing mostly various heads, like of people, including mine.End ends with a "to be continued" slide and I could definitely see having your picture as a part of my matrix.KirbyOn Mon, Feb 17, 2025 at 7:41 PM kirby urner <kirby...@gmail.com> wrote:I'd be up for a Tom's breakfast.I like the idea of an ISEPP function involving Keen but don't think there's any rush to get there, as there's plenty to discuss in the meantime.My little world of Buckynuts is heating up in terms of more players engaging in inter-communications. Means more Zoom calls for me.Kirby
On Mon, Feb 17, 2025 at 7:32 PM Terry Bristol <terr...@gmail.com> wrote:
Kirby –I just did an ‘Interview’ featuring Adrian Bejan and me.Interesting that this guy Pedro in Spain set it up. All done wire Google Meet.Bejan accepted but insisted on me as his accomplice.Just received the link from Pedro: https://www.youtube.com/@IAISDIALOGUES/videosISEPP hasn’t done this in the past but we could invite Keen to make a video presentation.You and I would be commentators. And then we could allow audience questions.People love to be invited to strut their stuff.With the ISEPP background, I am guessing that Keen would accept.This could be the first of a series.Here is the link Pedro gave us to see what had gone on before.Let’s do a breakfast and catch up. I am liking your ‘ultimate’ conclusion that ‘it is all language’.I told Bjan that this is the natural extension of his work. I will explain.Terry=========================
Reply all
Reply to author
Forward
0 new messages