AI Companionship

 Expected.

AI Companionship

“The great thing about AI is that it is constantly evolving. One day it will be better than a real [girlfriend]. One day, the real one will be the inferior choice.” 

This quote is from one of more than 20,000 users who joined the early community of CarynAI, a voice chatbot created by influencer Caryn Marjorie. Users pay $1 per minute to talk to an AI version of Caryn—your “virtual girlfriend”—and she made $72,000 in the first week. CarynAI is just one of dozens of ways that consumers have been forming real relationships with artificial intelligence.

Having an AI companion might seem niche, but it’s emerged as a predominant use case for generative AI. There are already hundreds of thousands, if not millions, of people—including us—who have already built and nurtured relationships with chatbots. We believe we’re on the cusp of a significant societal shift: AI companions will soon become commonplace. What was once the domain of science fiction, the likes of WALL-E, R2-D2, or Plankton’s Karen, is fast becoming a reality. And at a16z, we are excited about it.

Many of the popular use cases today are romantic, which won’t be surprising to anyone who has studied the history of the consumer internet. The most sophisticated consumers of generative models today are hacking together virtual companions, coordinating their efforts across platforms like Reddit, Discord, and 4chan to find tools and evade censors. In fact, we know of underground companion-hosting services that are globally distributed, with tens of thousands of users. This is an early developer community that we take seriously—they may be fringe today, but they’re building tomorrow’s mainstream products.

We’re also starting to see early signs of a much broader set of applications beyond AI “boyfriends” and “girlfriends.” Snapchat recently revealed that 150 million people sent more than 10 billion messages to its chatbot in its first two months. What were they chatting about? It turns out pets, pop culture news, and soccer were some of the top conversation topics.

Yes, chatbots have been around for decades, but we believe that this time really is different. Today’s bots aren’t merely a step function improvement on 1:1 conversations—they’re making inroads into our social lives. AI companions are seamlessly blending into our relationships with friends and family members, and they’re joining our communities like any other human.

In this post, we’ll dive into the emerging behavior of AI companionship: why people are creating companions, how they’re doing it, and what features or use cases might be coming next.... ' 

Looking Beyond LLM

Looking beyond, fascinating

Emerging Architectures for LLM Applications

by Matt Bornstein and Rajko Radovanovic

AI, machine & deep learning  enterprise & SaaS  AI  Generative AI  machine learning

TABLE OF CONTENTS

The stack

Design pattern: In-context learning

Data preprocessing / embedding

Prompt construction/ retrieval

Prompt execution / inference

What about agents?

Looking ahead

Explore more: AI + a16z

Large language models are a powerful new primitive for building software. But since they are so new—and behave so differently from normal computing resources—it’s not always obvious how to use them.

In this post, we’re sharing a reference architecture for the emerging LLM app stack. It shows the most common systems, tools, and design patterns we’ve seen used by AI startups and sophisticated tech companies. This stack is still very early and may change substantially as the underlying technology advances, but we hope it will be a useful reference for developers working with LLMs now.

This work is based on conversations with AI startup founders and engineers. We relied especially on input from: Ted Benson, Harrison Chase, Ben Firshman, Ali Ghodsi, Raza Habib, Andrej Karpathy, Greg Kogan, Jerry Liu, Moin Nadeem, Diego Oppenheimer, Shreya Rajpal, Ion Stoica, Dennis Xu, Matei Zaharia, and Jared Zoneraich. Thank you for your help!

The stack

Here’s our current view of the LLM app stack (click to enlarge): ... '

Generative AI is Coming for Insurance (May 2023 Fintech Newsletter) with GPT Definitions.

A good indicator is usually  how investment or insurance reacts.

Generative AI is Coming for Insurance (May 2023 Fintech Newsletter)

by Joe Schmidt, Sumeet Singh, and Seema Amble

fintech  Generative AI

By Andreessen Horowitz

TABLE OF CONTENTS

Generative AI is Coming for Insurance

Opportunities & Risks with Third-Party Payment Links

Visa+, Interoperability, and Creating Clearinghouses for New Payment Methods

Featured Tweetstorms

More From the Fintech Team ...  

Generative AI is Coming for Insurance

Because underwriting, selling, and servicing rely so heavily on humans processing large quantities of written or verbal communication, existing tools have struggled to properly automate these services and materially impact loss ratios (losses on written premiums) and expense ratios (underwriting and servicing written premiums). Large language models (LLMs), with their ability to proficiently collect and distill large amounts of data, could change this as they can augment or fully replace the process of a human combing through large amounts of data. 

While current machine learning technology allows for improved decisioning on simple products like auto and home insurance, more complex underwriting processes like commercial and life insurance remain challenging. This has less to do with the process of decisioning relevant data and more to do with collecting and synthesizing the relevant data. While traditional ML models have helped dramatically improve more standardized underwriting processes like home and auto, LLMs could potentially help with the more complex group by gathering data to help underwriters make better decisions, especially in more intricate cases like large commercial policies where more context and follow-up questions are required. For example, most large commercial policies cover dozens or more locations, and each location has specific nuances (such as electrical panels, fire doors, sprinkler density/effectiveness, management effectiveness, amount of combustible storage) that must be gathered from the applicant, understood by the underwriter, and evaluated against underwriting guidelines. LLM-powered workflow software for underwriters could drive down underwriting time and cost while increasing accuracy.

On the sales side, considered purchases, like life or disability insurance and annuities, are primarily sold offline through human agents and brokers because they’re complicated products that buyers often have questions about. (Consumers are quicker to buy mandatory insurance products, like home or auto insurance, online.) LLMs trained on customer data or materials on what policies are appropriate for a certain customer situation could help answer complex questions for consumers about what policies they should buy and how that policy might impact their unique needs.

And finally, carriers and agencies employ large policy-servicing divisions to help with changing policies, customer support, and claims, as well as “internal wholesaler” teams to constantly monitor and service the production of affiliated agencies or brokerages. Think of these as vertical-specific call centers where a representative needs to distill what a customer, agent, or broker actually needs during a conversational dialogue, and either respond with the answer or enter the appropriate information into a system. Allowing LLMs to manage some of these conversations could dramatically improve efficiency and profitability. ... 

—Joe Schmidt, a16z fintech partner  (And More) 

Using Generative AI to Unlock Probabilistic Products

Interesting idea ....

Using Generative AI to Unlock Probabilistic Products    in a16z  in Andreessen 

by Anish Acharya

Enterprise & SaaS  fintech  product management  Generative AI

One remarkable thing about the invention of generative AI is that it’s one of the first examples of a probabilistic computer that produces outputs that are varied and non-deterministic. Some, who expect these systems to behave like traditional systems, have complained about hallucinations—however, these complaints miss the point. A dispersion of outputs (including hallucinations) is exactly what’s so special here. In fact, they unlock a whole new category of product design: probabilistic products.

By probabilistic products, we mean products with non-deterministic and often emergent attributes. Social networks are an example of probabilistic products. The emergent behaviors that appear on these networks typically cannot be predicted but, instead, can simply be observed. One way to think about this is to organize product categories, or a product’s use cases, into three groups: '

1 Those that are uniquely enabled by or benefit from the probabilistic nature of the platform

2 Those that are tolerant of the probabilistic nature of the platform

3 Those that are intolerant of the probabilistic nature of the platform  .... '

Made Me Think: About SynthAI

Still thinking this, Like to test its usefulness, efficiency.  Join me.

For B2B Generative AI Apps, Is Less More?   by Zeya Yang and Kristina Shen in Andreessen Horowitz

AI, machine & deep learning  enterprise & SaaS  Generative AI

Table of contents

Wave 1: Crossing the bridge from consumer to enterprise

What’s the cost (or benefit) of disrupting the workflow?

Wave 2: Converging information for improved decision making

Implementing SynthAI

A battle to own the workflow

We’ve watched large language models (LLMs) become mainstream over the past few years and have studied the implementations in the context of B2B applications. Despite some enormous technological advances and the presence of LLMs in the general zeitgeist, we believe we’re still only in the first wave of generative AI applications for B2B use cases. As companies nail down use cases and seek to build moats around their products, we expect a shift in approach and objectives from the current “Wave 1”  to a more focused “Wave 2.”

Here’s what we mean: To date, generative AI applications have overwhelmingly focused on the divergence of information. That is, they create new content based on a set of instructions. In Wave 2, we believe we will see more applications of AI to converge information. That is, they will show us less content by synthesizing the information available. Aptly, we refer to Wave 2 as synthesis AI (“SynthAI”) to contrast with Wave 1. While Wave 1 has created some value at the application layer, we believe Wave 2 will bring a step function change.

Ultimately, as we explain below, the battle among B2B solutions will be less focused on dazzling AI capabilities, and more focused on how these capabilities will help companies own (or redefine) valuable enterprise workflows. ... '      (charts at the link at Andreessen)

Interesting Auction Tech

Worked on some auction tech way back when, so this interested me.  Not to say I fully understand it, but like to see new and differ ways of valuing  things.  

Through the looking glass: A cross-chain sealed-bid auction using Aztec Connect,    Michael Zhu

AUCTION DESIGN, TABLE OF CONTENTS

How it works, Hiding bids via “splitting”

Aztec Connect for anonymous bidding

InputValue

How it compares, Gas costs, User experience, Privacy

Editor’s note: This piece is part of our ongoing series on all things auctions for web3. Part 1 was an overview of technical challenges (and opportunities) specific to designing on-chain auctions. Part 2 was a piece on clearing the market and avoiding gas wars. Part 3 and Part 4 explore how auction theory translates into practice by introducing two implementations of sealed-bid auctions. 

Over the course of this series, we’ve explored different strategies for bridging the gap between auction theory and what can be built on-chain, each with its unique nuances. In particular, we’ve focused on implementing sealed-bid, second-price (Vickrey) auctions, which have been used for decades in sales of art, timber, and ad space. But we rarely see them implemented trustlessly using smart contracts, in part due to the difficulty of implementing private bids. The transparent nature of public blockchains can prevent dishonest intermediaries from censoring bids or manipulating auction results; but they also impose challenges on developers, who must find novel solutions for protecting their users’ privacy. 

In our previous auction implementations, we’ve used two approaches to keeping on-chain bids private: The first (OverCollateralizedAuction) conceals bid values using overcollateralization (where bidders lock up more collateral than required by the bid), protecting privacy at the expense of capital efficiency. This led us to our second design (SneakyAuction), which uses the CREATE2 opcode to camouflage bids among other transfers on the blockchain. But empirical analysis showed that this approach would not be effective at hiding very large bids that deviate from amounts typically transacted on Ethereum. 

In this post, we return with a novel cross-chain approach (called AztecConnectAuction) that can provide privacy to bids of any size, without requiring extra collateral. Relying on multiple blockchains achieves functionality that wouldn’t be possible on any single chain alone, and provides a different set of benefits and tradeoffs for developers to consider when implementing their own protocols. In our case, we use Aztec Connect to leverage the anonymity of Aztec’s ZK rollup, while retaining the benefits of Ethereum L1 settlement. 

We’ve added the implementation to our Auction Zoo repository on GitHub, alongside our previous auction implementations. We hope you’ll build on these, share your ideas, and send us your feedback.

How it works

First a quick refresher on the Vickrey auction: Bidders submit private bids (traditionally for a single item) to the auctioneer in sealed envelopes. The highest bidder wins, but pays the second-highest bid. To translate these characteristics on-chain, our new auction follows the same blueprint as before: Bids are collateralized and committed to during the “bidding phase”, in such a way that doesn’t reveal their precise values until they are revealed in the subsequent “reveal phase”. Once the reveal phase is over, the auction can be ended and the winning bidder pays the seller in the amount of the second-highest bid (the Vickrey payment rule). Our auction implementations differ in how they keep bids hidden during the bidding phase — this time, we will use a mechanism that is capital efficient even for large bids. 

Hiding bids via “splitting”

In OverCollateralizedAuction, privacy relied on overcollateralization to hide the exact bid value. In SneakyAuction, we were able to achieve bid privacy by using the CREATE2 opcode to hide the intent of a bid transaction — the value of the transaction is publicly visible, but the transaction is indistinguishable from a normal ETH transfer.

An alternative approach is to obfuscate the value of a bid by splitting it into multiple, unlinkable transactions. Imagine an auctioneer receives three checks for $100, $200, and $300, each signed with invisible ink. All three could have been signed by the same bidder, tendering a bid for $600. Or, they could be three separate bids, signed by three respective bidders. Or, there might be two bidders, one of whom has split their bid into two checks. Even if the check amounts are public, the underlying bid values are hidden until the signatures are revealed.   .... ' 

Five Predictions for the Future of Learning in the Age of AI

Good, thoughtful piece ...

Five Predictions for the Future of Learning in the Age of AI      by Anne Lee Skates  in A16Z

AI, machine & deep learning  education  Education  Generative AI

When OpenAI released its chatbot ChatGPT last year, proponents were quick to announce the death of various writing-related fields, such as screenwriting, computer programming, and music composition. One particular field stood out as a sector that would feel the power of ChatGPT almost immediately: education. With ChatGPT’s technology, students can now easily cheat on papers and college admissions essays, while on the opposite end, teachers can outsource their curriculums to AI—and no one would be the wiser. 

But ChatGPT is hardly the end of education. Just as quickly as students started passing off the chatbot’s work as their own, new programs popped up to detect AI-written work, and teachers, looking to get ahead of their students, started integrating ChatGPT responses into their lesson planning. 

The truth is, if leveraged well, AI has the potential to greatly enhance students’ abilities to think critically and expand their soft skills. And for skeptics who are worried kids will stop learning basic skills, avoid practicing, and forget general facts if they can rely on an AI to answer for them, psychologists Edward Deci and Richard Ryan posit in their self-determination theory that humans are intrinsically driven by autonomy, relatedness, and competence—that is, they will continue to learn regardless of any shortcuts thrown their way. The creation of Wikipedia is a great example. We didn’t stop learning history or science just because we could now quickly look up dates and formulas online. Instead, we simply gained an additional resource to help us fact-check and facilitate learning.

Seeing as education is one of AI’s first consumer use cases, and programs like ChatGPT are how millions of kids, teachers, and administrators will be introduced to AI, it is critical that we pay attention to the applications of AI and its implications for our lives. Below, we explore five predictions for AI and the future of learning, knowledge, and education.

1. The one-on-one model goes mainstream 

Getting one-on-one support for services like tutoring, coaching, mentorship, and even therapy was once only available to the well-off. AI will help democratize these services for wider audiences. In fact, Bloom’s 2 sigma problem—which found that students who received one-on-one teaching performed two standard deviations better than children in a traditional classroom—has a solution now. AI can potentially act as a live tutor for anyone, with humans supplementing the AI to provide in-depth knowledge and emotional and behavioral support. Academic tool Numerade, for example, recently released an AI tutor, Ace, that can generate personalized study plans, curating the right content depending on students’ skill levels.   .....   (four more predictions follow) 

Generative AI and Games

Intriguing approach .

The Generative AI Revolution in Games

Excerpt from James Gwertzman and Jack Soslow   in a16z.com,  Andreessen Horowitz

AI, machine & deep learning  gaming, social, and new media  Generative AI  machine learning

To understand how radically gaming is about to be transformed by Generative AI, look no further than this recent Twitter post by @emmanuel_2m. In this post he explores using Stable Diffusion + Dreambooth, popular 2D Generative AI models, to generate images of potions for a hypothetical game.

What’s transformative about this work is not just that it saves time and money while also delivering quality – thus smashing the classic “you can only have two of cost, quality, or speed” triangle. Artists are now creating high-quality images in a matter of hours that would otherwise take weeks to generate by hand. What’s truly transformative is that:

This creative power is now available to anyone who can learn a few simple tools.

These tools can create an endless number of variations in a highly iterative way.

Once trained, the process is real-time – results are available near instantaneously.

There hasn’t been a technology this revolutionary for gaming since real-time 3D. Spend any time at all talking to game creators, and the sense of excitement and wonder is palpable. So where is this technology going? And how will it transform gaming? First, though, let’s review what is Generative AI?

Generative AI is a category of machine learning where computers can generate original new content in response to prompts from the user. Today text and images are the most mature applications of this technology, but there is work underway in virtually every creative domain, from animation, to sound effects, to music, to even creating virtual characters with fully fleshed out personalities.

AI is nothing new in games, of course. Even early games, like Atari’s Pong, had computer-controlled opponents to challenge the player. These virtual foes, however, were not running AI as we know it today. They were simply scripted procedures crafted by game designers. They simulated an artificially intelligent opponent, but they couldn’t learn, and they were only as good as the programmers who built them.

What’s different now is the amount of computing power available, thanks to faster microprocessors and the cloud. With this power, it’s possible to build large neural networks that can identify patterns and representations in highly complex domains.

This blog post has two parts:

Part I consists of our observations and predictions for the field of Generative AI for games. ... ' 

Can't be Evil NFT Licenses?

Admit I do not entirely understand, reconsidering: 

The Can’t Be Evil NFT Licenses  Formerly from A16xcryprom,  Andreessen 

by Miles Jennings and Chris Dixon, August 31, 2022

Two decades ago, the newly formed Creative Commons (CC) released its first set of free, public licenses, enabling creators to open up aspects of their copyrighted work to the public for sharing, remixing, and reuse beyond the default “all rights reserved” notice. Today more than two billion CC-licensed works exist — among them the popular xkcd webcomics by Randall Munroe; user-generated content sites like Flickr; open access to digital images of public-domain artworks displayed in New York’s Metropolitan Museum of Art; the online science journal PLOS One; and educational resources like Khan Academy and Wikipedia.

A key feature of the Creative Commons model is levels of permissions granted by the original creators or copyright holders — whether for adaptations, derivatives, commercial use, and so on — with CC0 being the most permissive as it essentially dedicates the copyrights to the public domain. Previous copyright licensing regimes were overly restrictive for many creators, and couldn’t keep pace with what the internet and then-new digital technologies made possible. This limited creators and the larger community from participating in shared “culture and knowledge production,” a movement that is only growing in importance today. 

Now that web3 innovations are testing the limits of traditional legal frameworks, it’s time for a new set of licenses, designed specifically for non-fungible tokens, or NFTs. The recent wave of CC0 (no-rights reserved) NFT projects, for example, has spotlighted the Creative Common’s most permissive agreement, but prominent creators (including record-smashing graphic artist Beeple) have used some form of CC license for years, while other NFT projects choose different customized terms. However, many NFT projects omit licenses altogether, or draft licenses that create more ambiguity than they resolve. Some copyright vulnerabilities have led to significant confusion around NFT licenses, and a number of other legal problems. 

To help address these issues, we’re releasing a set of free, public “Can’t Be Evil” Licenses, designed specifically for NFTs and inspired by the work of Creative Commons. The licenses are freely available for use by the community, and serve three goals: (1) to help NFT creators protect (or release) their intellectual property (IP) rights; (2) to grant NFT holders a baseline of rights that are irrevocable, enforceable, and easy to understand; and (3) to help creators, holders, and their communities unleash the creative and economic potential of their projects with a clear understanding of the IP framework in which they can work. Since most early-stage projects don’t have access to legal resources, we worked with some of the foremost IP lawyers in the web3 space to design six types of broadly applicable NFT licenses and make them available for all. 

The case for NFT-specific licenses

Many people buy NFTs to own an avatar, an artwork, or any number of other creative outputs — but the reality is they usually can’t be sure of what they’re getting. When you buy an NFT today, you’re usually purchasing a tokenID (stored on a blockchain), along with metadata that “points” or refers to some other content file (typically stored off-chain, though there are examples of fully on-chain artwork). This fact causes confusion regarding rights of NFT buyers in the vast majority of cases.

US copyright laws do not automatically grant buyers of artwork (both traditional and digital works) the right to reproduce, adapt, or even publicly display the artwork. Without a license or assignment of the copyright from the NFT creator, the buyer cannot exercise any of the rights under copyright (such as reproduction, adaptation, and public display) except through copyright exceptions such as “fair use” which are narrow and uncertain. 

Licenses allow creators to grant holders additional rights, but to date, licenses aren’t applied consistently across projects. Many projects launch without licenses, or with custom licenses that create more ambiguity than they resolve. Licenses (and other documentation of what buyers are legally allowed to do with their NFTs) are often kept off-chain, where they could be changed in ways holders don’t expect. 

These issues are compounded by the fact that copyrights are notoriously difficult to transfer. Even a savvy buyer has no way to inspect an infinite bundle of rights and know which ones a previous owner may have already given away. 

Standardized NFT-specific licenses should ideally be tracked and enforced on the blockchain to provide more certainty for users. Better licensing frameworks have the potential to make high quality licenses more readily available, clear up ambiguity around ownership, and save creators some of the burden (and expense) of creating their own licensing regimes. 

Applying the “Can’t Be Evil” principle to NFT licenses: .... 

Building Reliable AI Models

 Below quite interesting,  here just the intro, full article linked to. Reviewing.  I note the point about changing data in context, which I always emphasize in models. 

7 Techniques for Building Reliable AI Models  By Beena Ammanath in future.16z.com

This is an edited excerpt from Trustworthy AI: A Business Guide for Navigating Trust and Ethics in AI by Beena Ammanath (Wiley, March 2022). Ammanath is executive director of the Global Deloitte AI Institute and leads Trustworthy & Ethical Technology at Deloitte. She has held leadership positions in artificial intelligence and data science at multiple companies, and is the founder of Humans For AI, an organization dedicated to increasing diversity in AI.

With AI model training, datasets are a proxy for the real world. Models are trained on one dataset and tested against another, and if the results are similar, there is an expectation that the model functions can translate to the operational environment. What works in the lab should work consistently in the real world, but for how long? Perfect operating scenarios are rare in AI, and real-world data is messy and complex. This has led to what leading AI researcher Andrew Ng called a “proof-of-concept-to-production gap,” where models train as desired but fail once they are deployed. It is partly a problem of robustness and reliability.

When outputs are inconsistently accurate and become worse over time, the result is uncertainty. Data scientists are challenged to build provably robust, consistently accurate AI models in the face of changing real-world data. In the information flux, the algorithm can meander away, with small changes in input cascading into large shifts in function.

To be sure, not all tools operate in environments prone to dramatic change, and not all AI models present the same levels of risk and consequence if they become inaccurate or undependable. The task for enterprises as they grow their AI footprint is to weigh robustness and reliability as a component of their AI strategy and align the processes, people, and technologies that can manage and correct for errors in a dynamic environment. .... ' 

Boutique Search Again

Reminds me of the early days of search and things we set up for specialty use.   Especially healthcare based, but could be related to any domain.    Curated web and supporting search engines.    Recall setting up a search engine pointed (curated to) directly at a specific project.   See 'Zakta' as an example, is it still available? 

The Future of Search Is Boutique, By Sari Azout in future.a16z

This is an edited version of a post that originally ran here. 

For most queries, Google search is pretty underwhelming these days. Google is great at answering questions with an objective answer, like “# of billionaires in the world” or “What is the population of Iceland?” It’s pretty bad at answering questions that require judgment and context like “What do NFT collectors think about NFTs?”

The evidence is everywhere. These days, I find myself suppressing the garbage Internet by searching on Google for “Substack + future of learning” to find the best takes on education. We hack Twitter with the “what is the best” posts over and over again. When I’m researching a new product, I type “X item reddit” into Google. I find enormous value in small, niche, often forgotten sites like Spaghetti Directory.

There’s an emergence of tools like Notion, Airtable, and Readwise where people are aggregating content and resources, reviving the curated web. But at the moment these are mostly solo affairs — hidden in private or semi-private corners of the Internet, fragmented, poorly indexed, and unavailable for public use. We haven’t figured out how to make them multiplayer. In cases where we’ve made them public and collaborative — here is a great example — these projects are often short-lived and poorly maintained.

The stated mission of a company worth almost two trillion dollars is to “organize the world’s information” and yet the Internet remains poorly organized. Or, stated differently, in a world of infinite information, it’s no longer enough to organize the world’s information. It becomes important to organize the world’s trustworthy information.  .... ' 

On Augmenting Intelligence with AI

 Interesting piece.  How do you fit in  ' AI' components into an intelligent assistant framework?  Do we have the needed pieces?  What else need to be built and enhanced.

AlphaFold, GPT-3 and How to Augment Intelligence with AI

By Niko Grupen  in A16z  Andreeseen

This is the first post in a two-part series. Read Part 2 here. 

Around the same time that Alan Turing was shaping his theories of machine intelligence in Manchester, another future giant of the computing world, Douglas Engelbart, was developing an alternative computing paradigm over 5,000 miles away in the Bay Area. 

Engelbart believed that computers, with their ability to synthesize and manipulate vast quantities of information, should help humans solve problems, rather than remove them from the problem-solving loop. This ideology is now known as augmented intelligence. Engelbart’s contributions to the field (both as a PhD student at UC Berkeley and at SRI in the decades after) were perhaps best exemplified through “The Mother of All Demos” in 1968, where he unveiled for the first time many of the computing features we now take for granted — the mouse, GUIs, hyperlinks, word processing, version control, and even video conferencing — in a single demonstration.

Although it’s enticing to think about artificial intelligence passing human equivalency tests like Turing’s Imitation Game (or maybe something more sophisticated for today’s generalist AI models), we really should be thinking about how Engelbart’s ideas translate to our modern AI era. Put another way, how do we build the next Mother of All Demos?

In reality, the next mother of all demos will be much more than just a demonstration. We already have the ingredients — a whole new set of AI tools — so now we need to think about how these ingredients can help us reimagine and redesign our existing workflows and user experiences. In doing so, we can usher in a new class of AI-native experiences for search, scientific research, game design, and more.

The Mother of All Demos in the age of deep learning

If we’re building a version of the Mother of All Demos in the deep learning era, we could begin with these models and tools. Although research labs release new models seemingly every week, these are a great starting point to begin rethinking how we interact with technology:

GPT-3

The latest in OpenAI’s GPT series, GPT-3 is a 175-billion parameter language model that is trained on practically all of the text that exists on the Internet. Once trained, GPT-3 can generate coherent text for any topic (even in the style of particular writers or authors), summarize passages of text, and translate text into different languages. OpenAI also recently released a follow-up, InstructGPT, that incorporates human feedback to reduce harmful or biased outputs.

Copilot

Github’s Copilot takes “translating text into different languages” to the programming world. Built on top of OpenAI’s Codex — think GPT-3, but trained on words and code — Copilot is an “AI pair programmer” that will generate anything from individual lines to whole functions of code, based on a docstring describing what the code is supposed to do. Recently, DeepMind released a competing code-writing model, AlphaCode, that solves coding challenges from popular programming competitions.  .... ' 

Non Fungible Tokens

I heard about NFTs perhaps a year ago.  Someone I know was pushing investment in them.  NFT's are blockchain style encrypted 'tags' for unique digital things.  The tag acts as a certification of uniqueness.  See Wikipedias description:  https://en.wikipedia.org/wiki/Non-fungible_token  

I see that the fabled Andreessen Horowitz is investing in a marketplace of NFTs.

https://a16z.com/2021/04/02/nfts-readings-resources/      Their reading list (canon) about NFTs

https://www.coindesk.com/nft-marketplace-opensea-valued-at-1-5b-in-100m-funding-round-led-by-a16z     - Their investment in a marketplace of  NFTs. 

Mean anything?

Next Wave of Computing Innovation will be Driven by Crypto

 Andreessen Says it will all be about Crypto

Andreessen Horowitz  A16z

Crypto Fund III  by Chris Dixon, Katie Haun, and Ali Yahya

cryptocurrencies & blockchains  a16z crypto  announcements

We believe that the next wave of computing innovation will be driven by crypto. We are radically optimistic about crypto’s potential to restore trust and enable new kinds of governance where communities collectively make important decisions about how networks evolve, what behaviors are permitted, and how economic benefits are distributed. That’s why today we’re pleased to announce a new $2.2 billion fund to continue investing in crypto networks and the founders and teams building in this space. This represents the beginning of an exciting new chapter for the a16z crypto team.

This fund allows us to find the next generation of visionary crypto founders, and invest in the most exciting areas of crypto. We invest in all stages, from early seed-stage projects to fully developed later-stage networks.

Expanding our team to provide unrivaled regulatory and operational capabilities alongside our crypto-native data science and research services.

When we started a16z crypto, our ambition was to bring together a highly specialized and crypto-native team that had never existed in venture capital to support the companies and projects we invest in, help them navigate the complexities of this space, and help projects design crypto protocols and mechanisms. We’ll continue to deepen our focus on the kind of behaviors our entrepreneurs expect from a crypto investor including staking, delegating, running nodes, actively participating in governance and designing mechanisms. We’re expanding our data science and research teams to support this effort. 

As with any new computing movement, crypto has endured a variety of challenges and misconceptions. That’s why we are also bringing together heavy-hitters across several functions to help translate crypto to the mainstream. Our team now includes experts in marketing, public relations, policy, regulatory affairs, recruiting, as well as general startup management. We are promoting Anthony Albanese, who joined us last year from the New York Stock Exchange, to Chief Operating Officer to lead operations.  ....... ' 

Incomplete Contracts

Followup on previous piece: 

Incomplete Contracts  by Jesse Walden  in Andreessen Horowitz

cryptocurrencies & blockchains

One way to think about various kinds of crypto projects is through the lens of contract theory. An axiom of this area of legal scholarship states: “all but the simplest contracts are incomplete”. That is, contractual arrangements cannot anticipate every possible outcome or set of actions, given complex and dynamic changes in the world the contract lives in.

When contracts are incomplete, they rely on renegotiation when unexpected contingencies like bankruptcy, regulation or even simple changes in details emerge. Such contingencies often require third parties like the legal system to help interpret and mediate between the two parties, and can lead to unpredictable outcomes. In this way, contracts are always about the unknown eventualities of decision making, incentives, and governance authority.

But if we accept that contracts are simply decision logic — akin to computer programs, then contract theory gives us a framework for thinking about different types of smart contracts and crypto-enabled projects — and how they can scale (including governance of them) .... ' 

From Blockchain to Contracts

Have been looking at the use of 'smart contracts' in a broader way.  Is this one direction?   Incorporating trust.

Investing in Aleo  by Katie Haun and Ali Yahya  in Andreessen Horowitz

From the beginning, our core thesis has been that the best way to think of a modern blockchain is as a new class of computer that has the ability to run a special kind of program. These programs are sometimes called smart contracts, and they’re different from ordinary programs in that they have a life of their own. They are independent, and once written, they obediently execute themselves subject to nobody’s authority. Because of this property, smart contracts are uniquely capable of earning trust. 

But smart contracts today have two big limitations: (1) they are fully transparent by design and therefore don’t allow for privacy (2) they don’t scale to millions (let alone billions) of users. These limitations exist because trust requires verification. Transactions on a blockchain need to be transparent so that everyone can verify that they are correct. And, they tend not to scale because it takes time and energy for all computers on the network to perform that verification.

But research in a cutting edge area of cryptography called zero-knowledge proofs promises to unlock an elegant solution to the privacy and scalability problems. We spent a great deal of time looking at various approaches and teams working on this. Aleo’s solution is both elegant and pragmatic.  ... '

Investing in Eco

How can more people become involved in Crypto?

Investing in Eco, Crypto

by Katie Haun and Arianna Simpson   in Andreessen Horowitz

One of the challenges in any new industry — crypto being no exception — is building things that are not just cool for the sake of cool, but that manage to reach and delight a broad set of users. Technology is at its best when it’s improving the lives of people in tangible, concrete ways. And with new technology like crypto, the strongest approach is often to give users those benefits without requiring them to get too involved in the nuts and bolts.

At a16z Crypto, we are constantly on the lookout for paths to get cryptocurrency into the hands of the next billion people. How do we think that will happen? By helping them achieve what they already want to do: spend, save, and make money — and by focusing users on tangible benefits, not on the underlying technology. 

That’s why we’re excited to announce our investment in Eco. 

Eco’s mission is to reimagine not only how consumer financial products work — but also to reimagine money itself. Eco is not a bank, checking account, or credit card. Although it has elements of each, it’s better than all of those. It begins as a single wallet that rewards you anytime you save or spend your dollars. 

Today, financial transactions involve numerous players at every level — banks, credit cards, processors, and more — introducing inefficiencies and fees along the way. Without the burden of legacy, misaligned business models, Eco is able to offer rewards that sound almost too good to be true, but aren’t. 

As investors, we’re in the business of finding and backing big ideas, and Eco has no shortage of those. The challenge is often finding teams who have the skills to back up that vision and bring it to life. 

In Eco, we found both.   ... ' 

Crypto Canon: From Andreessen Horowitz

Very nicely done resource here,  with lots of introductory and related information.

See the Andreessen Crypto Canon:  https://a16z.com/2018/02/10/crypto-readings-resources/    

From Andreessen Horowitz  Crypto Canon  Here’s a list of crypto readings and resources. It’s organized from building blocks and basics; foundations (& history); and key concepts — followed by specific topics such as governance; privacy and security; scaling; consensus and governance; cryptoeconomics, cryptoassets, and investing; ….   Feb 10th, 2018

More generally their Crypto related writings and ongoing coverage are in blog form here: https://a16z.com/crypto/ 

The Social Clubhouse

Writeup on 'Clubhouse', mentioned to me in a conversation.    As a possible next big social media thing.   Note the Horowitz support.  As usual,  when I read these things I always end up thinking 'whats the use of that'?    But I have been wrong.

The fatal flaw of Clubhouse   By Mark Schaefer

 Clubhouse appears to be the next hot social media platform.

Backed by the prestigious venture capital firm Andreessen Horowitz and growing faster than Mark Zuckerberg’s bank account, Clubhouse is a platform where users participate in different chat rooms on a wide range of topics. The conversations are audio-only, and when they finish, they disappear forever. It’s like talk radio on steroids — endless, diverse, inclusive. Wading through Clubhouse can be dizzying because there are so many interesting conversations.

Clubhouse is halfway between a podcast and a party. It’s a facilitated discussion, no make-up required. This is why it will displace at least some podcast consumption — you have a chance to actually participate with the hosts, and that’s special.  ... "

Doing Old Things Better Vs. Doing Brand New Things

Old Conundrum ....  But I think good to consider the positives and negatives of both independently.  'Doing things' vs investing also quite different, gets you different kinds of money, which depends upon the current economy.     And aso who you get to pay attention.   Very new things entice creators,  clever widgets inspire engineers.   

Doing Old Things Better Vs. Doing Brand New Things   by Chris Dixon in Andreessen

AI, machine & deep learning  cloud computing  cryptocurrencies & blockchains how innovation happens

New technologies enable activities that fall into one of two categories: 1) doing things you could already do but can now do better because they are faster, cheaper, easier, higher quality, etc. 2) doing brand new things that you simply couldn’t do before. Early in the development of new technologies, the first category tends to get more attention, but it’s the second that ends up having more impact on the world.... "