AI Unchained

Guy Swann

If you are looking for self-hosted and sovereign AI tools - you've come to the right place. Making sense of the technology and concepts behind AI, breaking down the tools to gain more freedom and autonomy in our lives, and investigating how to avoid being trapped in centralized systems of control. If you want a sovereign future in the digital age, this is where we will build it. This is AI Unchained.

Ai Take_037 - Focus

"Never half-ass two things. Whole-ass one thing."— Ron Swanson Focus. Easily spoken, hard to live by. There are many lessons that Bitcoin teaches, and one that is more clearly true the older I get, is that every single thing in a trade off. And not just solutions, or tools, or networks, but the very thing that you spend your time thinking about, at the expense of all other things. What dreams you spend your time building, at the expense of all other dreams. What new ideas you begin, at the expen

Nov 26, 2024 • 45:48

AI Take_036 - Does AI Even Have a Product?

Is the pursuit of artificial general intelligence a pipe dream? Are the economics of building massive AI models unsustainable? And what if the future of AI isn't about creating a single, all-powerful intelligence, but rather a multitude of small, specialized tools? In this episode, we explore the latest developments in AI and challenge the conventional wisdom about where the field is headed. Links to check out The Subprime AI Crisis Article (Link: https://tinyurl.com/39m3x2ws) Andrew McCarthy

Nov 11, 2024 • 50:36

AI Read_012 - The Subprime AI Crisis

What do numerous over-hyped product launches, massively subsidized compute, over extended investment, lack of customer adoption, unwillingness to pay extra, enormous and increasing costs, and total dependence on a few massive players have in common? The big tech Ai industry! Today we dive into a fantastic breakdown of the state of the ai market, and the potentially dismal financial forecast for the poster child of Ai success, OpenAi. Is there really a massive market here that supports 100 bi

Oct 10, 2024 • 1:01:40

Ai_035 - My Favorite Ai Tool

Today we dive into my favorite Ai app, combining dozens of open source tools unlocking a new era of creativity and productivity? How can you harness the power of AI to simplify workflows and organize our digital lives? Join me as I explore the possibilities and implications of AI-powered tools and share my own experiences with Pinokio and other innovative applications. Links to check out to explore more: Pinokio.computer (Link: Pinokio.computer) Comfy UI (Link: https://github.com/comfyanonymous

Sep 27, 2024 • 38:05

Ai_034 - Small Models Reign Supreme

Will the increasing shift towards smaller, more efficient AI models revolutionize the way we interact with AI? Can open-source models and fine-tunes like Hermes 3 and Grok 2 challenge the dominance of proprietary models? And what does the future hold for AI development as the hype and investment surrounding AI begin to wane? Join me as I explore these questions and more on this episode of Ai Unchained.Links to check out to explore more: Jimmy Song: AI vs Bitcoin (Link: https://primal.net/e/note

Sep 13, 2024 • 1:38:38

Ai_033 - Fear to Freedom: The Optimistic Case for AI [The Staying Free Podcast]

"When you introduce 2 million, 10 million "developers" that were never developers into an ecosystem that's used to having closed platforms and silos and stuff, what do those 10 million motivated people build? They're more likely to build the solution that they always wanted and that they build something that gets around these silos.One of the most important things that's going on right now is the shift from closed platforms to open protocols. The more that all of that stuff is open, the more the

Aug 23, 2024 • 2:00:48

Ai_032 - Open Source Ai is Catching Up

Open source Ai is catching up. In this episode of AI Unchained, we explore the implications of the dominant open source models actually competing with proprietary models. We challenge assumptions about AI development trajectories, drawing parallels with other technological evolutions. Could the next AI revolution be horizontal rather than vertical, and what might this mean for accessibility and innovation in the field? Links to check out Llama 3.1 and the 405B parameter model (Link: h

Aug 16, 2024 • 1:06:16

AI_031 - How I use AI

Is AI really all hype and oversold gimmicks, or is it going to change the world and replace everyone with software workers? Or are both of these views a distraction from what's right in front of us? How is AI useful today for very real tasks? In this episode, we'll dive into some major new open-source models and releases, along with exploring the very practical applications of LLMs with the help of Nicholas Carlini's fantastic article on where he finds value today. I also want to challenge the p

Aug 9, 2024 • 1:20:42

Ai_030 - Preparing for the Intelligence Explosion

How might open-source AI development shape the future of artificial intelligence and global competition? Guy Swan examines Leopold Aschenbrenner's "Situational Awareness" piece and Mark Zuckerberg's recent article on open-source AI, exploring the potential for an intelligence explosion, the risks of centralized AI development, and the advantages of a decentralized approach. Could open innovation be the key to maintaining technological leadership and ensuring a positive AI future? Links to

Aug 4, 2024 • 1:56:52

Ai Read_011 - The AGI Manhattan Project

"Do we need an AGI Manhattan Project? Slowly at first, then all at once, it will become clear: this is happening, things are going to get wild, this is the most important challenge for the national security of the United States since the invention of the atomic bomb. In one form or another, the national security state will get very heavily involved. The Project will be the necessary, indeed the only plausible, response." ~ Leopold Will the race for AGI lead to a Manhattan Project-like initiative

Jul 25, 2024 • 49:40

Ai Read_010 - SuperAlignment

"Reliably controlling AI systems much smarter than we are is an unsolved technical problem. And while it is a solvable problem, things could very easily go off the rails during a rapid intelligence explosion. Managing this will be extremely tense; failure could easily be catastrophic."~ Leopold Aschenbrenner As we approach a potential intelligence explosion and the birth of superintelligence, how can we ensure AI remains beneficial and aligned with the goals of furthering humanity, while naviga

Jul 18, 2024 • 1:45:40

Ai Read_009 - The Challenges of the Intelligence Explosion

What if the race to AGI isn't just about code, but a massive industrial mobilization? As AI capabilities grow exponentially, will we see trillion-dollar compute clusters and unprecedented energy demands? And in this high-stakes competition, how critical is securing AI secrets from rival nations? This episode delves into these questions, exploring the potential challenges and geopolitical implications of rapid AI advancement. Check out the original article by Leopold Aschenbrenner at situati

Jul 11, 2024 • 1:44:32

Ai_029 - New Tools and Major Moves in AI

Open AI signals that they are the surveillance leviathon we suspected they always were. Anthropic's Claude 3.5 may be even better than ChatGPT. I explore and detail out some new tools and build new tiny apps to recreate some of the incredible features of the Apple AI announcements, but in a local, open source way. A cybersecurity disaster in Microsoft's announcement of their new OS "feature" Recall and why it may be the worst thing for security and privacy in a very long time. And much more in t

Jul 4, 2024 • 1:33:00

Ai Read_008 - From AGI to Superintelligence

Are we on the verge of an intelligence explosion that could redefine the future of AI and humanity? Explore Leopold Ashen Brenner's insights on the leap from AGI to superintelligence and what it could mean for every aspect of our lives. Join me as we delve into the potential and perils of a superintelligent world and its profound implications. We are continuing a read situational awareness the decade ahead by Leopold Ashen Brenner that we started in the previous episode and this will be part two

Jun 27, 2024 • 1:41:34

Ai Read_007 - SITUATIONAL AWARENESS - The Decade Ahead

"Do not expect the vertiginous pace of progress to abate. The trend lines look innocent, but their implications are intense... Scaling up simple deep learning techniques has just worked. The models just want to learn, and we're about to do another 100,000X by the end of 2027. It won't be long before they're smarter than us."~ Leopold Aschenbrenner What does the next decade hold for AI, and are we on the brink of achieving artificial general intelligence? Join me as I dive into Leopold Ashen Bren

Jun 20, 2024 • 1:47:16

Ai_028 - How Much Do You Trust Apple?

Today Guy Swann dives deep into everything about the "Apple Intelligence" announcement, which aims to integrate powerful AI capabilities directly into the company's operating systems. He raises thought-provoking questions about the implications of such a system, particularly regarding privacy and trust in a tech giant like Apple. The episode promises to explore the potential benefits and drawbacks of this groundbreaking AI implementation. Links to check out: Ai_023 - The Promise of personal Ai

Jun 16, 2024 • 1:24:22

Ai_027 - Is AI Breaking the Web?

A trend I've noticed with Ai, is that it has shifted itself into the middleman position between everyone and the normal surfing of the web. While it may save the user time, could this be disrupting the very traffic that monetizes content on the internet? With this fundamental change in the monetization paradigm, a huge shift is coming to the web. What will it look like, and how should we prepare? Tune in to find out. Links to check out Descript (Link: https://tinyurl.com/bdhr75

Jun 6, 2024 • 54:12

Ai_026 - The Scam of Open AI

A very unique dynamic occurs with generative AI. LLM companies are scraping the internet for millions of works belonging to other people, photos, videos and film, forum posts, tutorials, all of this human created content meant to provide value to others in hopes of monetizing the attention, or gaining reputation from that research or creation.But without attribution or sharing of the proceeds, we can recreate, mimic, or even pull that exact content back out of the generative models. What does th

May 31, 2024 • 1:10:42

Ai_025 - Bullsh.t and Breakthroughs

Today we explore the deluge of announcements from both OpenAi and Google. With a plethora of Ai features dropping at Google I/O And Chat GPT-4o landing with an ai that can be spoken to like a human, how do we determine the difference between groundbreaking AI tools and mere gimmicks. How do we discern practical applications from overhyped features? Join Guy as he navigates the latest AI developments, asking the critical question: What truly enhances our digital lives and what falls short?

May 20, 2024 • 1:25:56

Ai_024 - CASCDR

"We're basically at a fork in the road. I think that there's a good chance we go the wrong way and you end up with a situation like China. We need to build these tools that serve Liberty and not tyranny. And we need to design these systems correctly. And then on top of that, I think, as I pointed out earlier, you know, we have an opportunity to build a better product than the tyrannical systems, you know, now before, before you're in too deep. So we need to do that so that peopl

May 9, 2024 • 1:26:23

Ai_023 - The Promise of Personal AI

How close are we to realizing the dream of personal AI? Dive into the potential of small, efficient AI models that enhance our self-awareness and improve our lives without compromising our privacy. Could these advancements lead us to a future where we not only manage but truly own our data, understanding ourselves better than any corporation could? Join us as we explore the fascinating possibilities of personal AI. Links to check out Atomic Habits (Link: https://tinyurl.com/hjcajxc) Faceboo

May 3, 2024 • 1:10:51

Ai_022 - Applying Ai

How can AI be harnessed to manage and enhance personal and professional digital interactions more effectively? In this episode of "AI Unchained," explore the essential tools and concepts needed to apply AI in self-hosted environments, navigate the intricacies of vector embeddings, and integrate AI seamlessly into your digital workspace. What modular approaches and AI skills are essential as coding becomes more automated? Dive into these questions and more with host Guy Swan. Links to

Apr 30, 2024 • 1:10:03

Ai_021 - Ai Hardware & the Next Wave

In today's episode of AI Unchained we dive into a plethora of new developments in Ai tailored hardware, the scope of what leaps in performance and capacity we may see in the next year, what it means for open source, how to access your local run Ai from anywhere with the Pear Stack, and yet another major tool in generative Ai that will take story telling to a new level? Join Guy Swan as he unpacks the latest and their profound implications for both developers and users alike. Links to check

Apr 19, 2024 • 1:01:39

Ai Read_006 - How (Actually) Open AI Wins

"As I see it, humanity is at a turning point where we must either embrace a small handful of totalitarian super intelligences or move AI to the edge, allowing a wide diversity of machine intelligences to flourish in cooperation with an equally diverse set of human intelligences, forming a truly decentralized hivemind. Fortunately, there’s mounting evidence that open source models will be the ultimate winners." ~Hivemind Ventures Our read from HiveMind Ventures today argues that the fu

Apr 5, 2024 • 1:24:55

Ai_20 - Instability Ai & Commoditized Intelligence

Today we delve into the leadership upheaval at Stability AI, Apple finally putting their cards on the table, the breadth of tools available for image and video generation, more ways to use LLMs to build micro apps, the release of numerous new models, and highlighting broader trends in the AI industry that may be rising to the surface. What might recent developments in the space reveal about the future of open source and centralized Ai? Are we headed toward a brighter, or much darker future? Find

Mar 30, 2024 • 1:25:26

Ai Read_005 - Here Lies the Internet, Murdered By Generative Ai

"All around the nation there are toddlers plunked down in front of iPads being subjected to synthetic runoff, deprived of human contact _even in the media they consume. There’s no other word but dystopian." ~ Mike Hoel What happens when the internet is so overrun with awful ai generated content, that it becomes too overwhelming to even find anything of value? And this is happening now, not in the future. What happens to a generation of toddlers who are watching incoherent, plotless, i

Mar 21, 2024 • 1:20:14

Ai_019 - What if You Had a Personal Software Developer

With the announcement of Devin, the "First AI Software Engineer" by Cognition Labs, some are calling it a breakthrough, while others aren't so impressed and believe other models can do largely the same thing. What is the unique difference that Devin brings to the table and does it have implications for the accessibility of coding and could it even suggest a change in direction for the future of software itself? Don't miss this fascinating topic in today's episode of Ai Unch

Mar 15, 2024 • 48:39

Ai_018 - Unlocking the world with AI

Advanced AI meets sustainable agriculture, opening a revolutionary pathway to food security. This episode brings Stephen Reisner from Potent Ponics into the spotlight, discussing a groundbreaking project that not only challenges the traditional AI applications in media but ventures into the heart of real-world, blue-collar issues. Explore how AI can empower farmers globally, breaking language barriers and democratizing access to organic cultivation knowledge. A blend of technology, agriculture,

Mar 7, 2024 • 1:36:40

Ai_017 - From Closed AI to Open Source with NVK

The only way to protect yourself from closed, surveilled AI tools, is to use the open and private alternatives. When NVK couldn't find them, he decided to build them. Unleashed.chat begins to realize the potentially for open source LLMs to compete with ChatGPT and make use of unique means to fine-tune and leverage vectorized data for higher accuracy and relevance to the tasks of the user. What goes into building a set of tools like this, and why are they so desperately needed? Find out in to

Feb 29, 2024 • 1:27:02

Ai_016 - Video AI Breakthroughs Already in 2024

"By the end of 2024, everyone will have tools easily available to them that will allow them to create any video they want by simply typing in a description and hitting generate." ~ Guy Swann Are we on the brink of a revolution where anyone can create blockbuster-quality videos from their bedroom? Discover how the latest AI breakthroughs in 2024 are dismantling the barriers to professional filmmaking and what this might mean for truth, politics, and storytelling. What will the world l

Feb 22, 2024 • 58:33

Ai_015 - Integration, Not Agents with Luke from ParsePrompt

We are diving into a discussion today with the creator of ParsePrompt on a perspective of where Ai can be leveraged for the highest multiplier. How to use Ai to make one form of content valuable over multiple mediums in an automated an repeatable fashion. As well as exploring the nuance of the open source alternatives, getting an LLM to speak with your data, and where hype has us chasing a foolish idea and missing what's right in front of us. Don't miss an awesome chat with Luke. Refere

Feb 10, 2024 • 1:37:59

Ai_014 - A Million New Devs with Hope TS

"I think AI can make a lot of things better, and if you put them together, if you are aware of the resources at your disposal, I think the average person can do more than the wealthiest person 200 years ago could have done" ~ HopeTS What if YOU could build the app you needed to solve your own problem? Ai has unleashed this ability to millions who have the knowledge necessary to tap into it, if only they understood how powerful this tool is. Building an app, specific to your exact need

Feb 2, 2024 • 1:22:32

Ai Read_004 - Building a Language Model (Part 2)

"it’s important to step outside of the hype and critically-analyze what is and is not useful. It’s very easy to get caught up in “potential” applications, and allow the imagination to go all exponential on you. It’s a very human thing. Turning that imagination into something tangible is what a business and an entrepreneur does. It’s our hope this report will be useful along that path." ~ Alek Svetski Today we dive into Part 2 of the Bitcoin & Ai Industry Report to learn the intri

Jan 25, 2024 • 2:03:37

AI Read #004 - Building a Language Model (Part 2)

"it’s important to step outside of the hype and critically-analyze what is and is not useful. It’s very easy to get caught up in “potential” applications, and allow the imagination to go all exponential on you. It’s a very human thing. Turning that imagination into something tangible is what a business and an entrepreneur does. It’s our hope this report will be useful along that path." ~ Alek Svetski Today we dive into Part 2 of the Bitcoin & Ai Industry Report to learn the intri

Jan 25, 2024 • 1:38:12

Ai Read_003 - The Nexus of Bitcoin and AI (Part 1)

There are so many myths and misconceptions around AI models. What they can do, how they are trained, even the fundamentals of how the data for them is even collected and formatted to make sense. The idea that we can "train our own model" by just feeding it a ton of conversation or notes is unrealistic, but why exactly is that? Today we get into a 2 Part read which will be one of the most value dense we have covered on the show yet. It breaks down the entirety of the process of identif

Jan 18, 2024 • 1:38:13

Ai_013 - Become Anyone With Ai

Today we explore the tools for voice changing ai and text-to-speech. What is Guy using, and what are the impacts of instant and easy voice copying be on content creation, copyright, the meme lords, fraud, and personal identity. Referenced in the show ElevenLabs.io (Link: https://elevenlabs.io/) Altered.ai (Link: https://www.altered.ai/) Episode #011 - The Fears of Ai (Link: http://tinyurl.com/4v9rj98r) Could Ai Go Nuclear? - Ai Unchained #003 - The Dark Side of Ai with Alex Lewin

Jan 11, 2024 • 52:29

Ai_012 - A Big Leap Forward for Local LLMs

The models, the tools, the implementations, everything is moving incredibly fast. But a very interesting trend that I've been seeing, is that the open source benchmarks are improving just slightly faster than the closed source alternatives. Slowly but surely, they are catching up. The day may come where the open source models are the best models, the ones with the greatest number of tools, and the largest variety of implementations, because naturally they're open and available for everyo

Jan 5, 2024 • 50:49

Ai_011 - The Fears of Ai

Ai is a powerful tool, and thus, necessarily comes with incredible risk. It could obsolete millions of workers, taking jobs and productive capacity away from humans and consolidate it into corporations. It could make the average individual too dangerous, putting incredible power into the hands of bad people. It may even present an existential threat to humanity, a veritable nuclear bomb, made available to everyone in the world... What could all of this mean? Doesn't it suggests we need to ti

Dec 22, 2023 • 1:04:13

Ai_010 - Tools for Running Local LLMs

Today we dive into a few of my favorite tools for running LLMs locally on your machine, even the average consumer machine. With a huge variety of models to choose from and varying levels of customization and presets, these will get you started on running your own Ai, and getting the pieces necessary to connect them into micro apps and other parts of your workflow. Not only are these huge productivity boosters to have, but it's also just a ton of fun to play around. For the large list of lin

Dec 14, 2023 • 39:47

Ai Read_002 - A Coder Considers The Waning Days of the Craft

"Coding has always felt to me like an endlessly deep and rich domain. Now I find myself wanting to write a eulogy for it." ~ James Somers Today we dive into an amazing piece and such an important perspective on the future with Ai. With GPT being able to code with ease, so many tasks and projects that would've taken hours before. Where the "so-so" programmer can now build nearly anything with an Ai partner, what is to come of the future of coding? Should we be celebrating

Dec 7, 2023 • 1:15:11

Ai_009 - 5 Ways You Should be Using Ai

If you are looking to maximize your productivity, to use the Ai tools available in a way that genuinely gives you that 5x improvement of your capabilities, this is the episode for you. These are the 5 ways I am currently using Ai more than any other to maximize my output. For other links mentioned in the episode: Ai transcription video from YT (Link: https://www.youtube.com/watch?v=ZDw8ZNFBKes) Premai.io (Link: https://www.premai.io/) gpt4all (Link: https://gpt4all.io/index.html) confyui

Nov 29, 2023 • 1:00:31

Ai Read_001 - A Gentle Introduction to Large Language Models

"Our first instinct when interacting with a Large Language Model should not be “wow these things must be really smart or really creative or really understanding”. Our first instinct should be “I’ve probably asked it to do something that it has seen bits and pieces of before”. That might mean it is still really useful, even if it isn’t “thinking really hard” or “doing some really sophisticated reasoning”. We don’t have to use anthropomorphization to understand what it is doing to provide us

Nov 17, 2023 • 1:47:57

Ai_008 - Ecash and Ai with Obi Nwosu

Get 50,000 free sats, plus sats back on everything you buy by signing up with Fold and the Spin+ Debit Card. ⁠⁠⁠⁠⁠⁠bitcoinaudible.com/fold⁠ Obi Nwosu, co-founder and CEO of Fedi, joins us today to talk about the new innovations and challenges in re-decentralizing the web and how both Bitcoin and Ai are at the core of making it possible. Guest Links Follow Obi on X (Link: https://x.com/obi) Follow Fedi on X (Link: https://twitter.com/fedimint) Fedi Website (Link: https://www.fedi.xyz/) H

Oct 9, 2023 • 1:20:03

Ai_007 - Global Markets for GPU Power with Christopher David

What if we could mitigate the enormous centralizing effects of the massive server farms and huge corporate entities, by pooling the GPU power of the millions of users and tinkerers who have unused or regularly idle computational power, into a decentralized marketplace for buying and selling Ai execution, and we could power it all with instantly settled sats over lightning? It just so happens that we can, and this ambitious project is being brought to life with GPUtopia. Chris David from the pro

Sep 30, 2023 • 1:09:01

Ai_006 - Evil, Intelligence & Value with Dhruv Bansal

Today we delve into a captivating topic, as we explore the potential relationship between AI and evil. My guest Dhruv Bansal brings his profound insights to the forefront as we navigate the depths of this subject. We contemplate the ethical dilemmas that arise when intelligence is harnessed for destructive purposes, and Dhruv expertly unpacks the intricate interplay between AI intelligence, good, and evil, leaving us with a lot to consider about its impact on our lives and the broader world. W

Aug 4, 2023 • 2:37:46

Ai_005 - Ai, Fiat Incompatible with Kody Low

"So we had built this for a while, but because API calls are basically free, there really wasn't a good match for it. There were some good matches for it and people would do it, but as they were doing it, it was like 'I can do it this way, but I can also do it with dollars.' This is kind of a really interesting paradigm when it comes to Ai stuff is that there's basically 3 or 4 different aspects of Ai that make it so you cannot use credit card relationships to solve the mone

Jul 4, 2023 • 1:27:03

Ai_004 - Bias, Ai's Best Feature with Aleks Svetski

Today we follow up from #002 with Alek Svetski to get him back on and digging into the open, community project they are releasing, the Spirit of Satoshi. They are exploring the capabilities of LLMs in combination with a data store in scaling one-on-one education that explores a topic or idea through a lens of explicit bias. Can we instill a variety of genuinely different worldviews and perspectives into Ai from which to test our understanding of the world? Svetski thinks we can. Hope you guys en

Jun 29, 2023 • 1:17:17

Ai Read_000 - Google "We Have No Moat..."

"These contributions were pivotal in the image generation space, setting Stable Diffusion on a different path from Dall-E. Having an open model led to ⁠product integrations⁠, ⁠marketplaces⁠, ⁠user interfaces⁠, and ⁠innovations⁠ that didn’t happen for Dall-E. The effect was palpable: ⁠rapid domination⁠ in terms of cultural impact vs the OpenAI solution, which became increasingly irrelevant. Whether the same thing will happen for LLMs remains to be seen, but the broad structural elements are

Jun 14, 2023 • 1:10:33

Ai_003 - The Dark Side of Ai with Alex Lewin

Ai may be able to make us vastly more productive, far more creative, and orders of magnitude more dangerous. As Ai moves into the realm of creating hierarchies of tasks to accomplish larger goals, combined with the unparalleled scalability of its execution, a broad, self-spreading and malicious Ai becomes frighteningly more possible by the day. Like a wildfire burning out of control, what could be the consequences, how bad could a hypothetical scenario get, and what, if anything, can we do about

Jun 10, 2023 • 1:34:08