A reminder for new readers. That Was The Week collects the best writing on critical issues in tech, startups, and venture capital. I select the articles because they are of interest. The selections often include things I disagree with. The articles are only snippets. Click on the headline to go to the original. I express my point of view in the editorial and the weekly video below.
This Week’s Video and Podcast
Content this week from : @gruber, @Knibbs, @sarahintampa, @stevesi, @finkd, @lexfridman, @pmarca ( ), @martinvars, @glichfield, @sama, @RazRazcle, @geneteare, @mgsiegler, @Scobleizer, @TimDraper
Contents
Editorial: Was I Wrong?
First Impressions of Vision Pro and VisionOS
Apple’s Vision Pro Isn’t the Future
Apple TV gets FaceTime, a new control center, a findable remote, and more
Sequoia spin-offs show US geopolitical games are a liability for Silicon Valley venture capital
Lex Fridman interviews Mark Zuckerberg (over 2 hours).
Why AI Will Save the World – Marc Andreessen
AI is a tool – Martin Varsavsky
Marc Andreessen Is (Mostly) Wrong This Time
OpenAI’s plans, according to Sam Altman
New Statement Signed by the Likes of OpenAI’s Sam Altman Warns of AI’s Extinction Risk
New Unicorns In May Hit Double Digits With Two AI Companies
Mercedes Just Beat Tesla to This Huge Milestone in the Race to Self-Driving Cars
GM EV owners will gain access to Tesla Superchargers in 2024
Sequoia Splits Its Global Fund Into 3 Separate Businesses
Twitter’s Ad Revenue Continues to Decline Under Elon Musk
SEC Files 13 Charges Against Binance, Founder Changpeng Zhao And Related Entities
Andrew Keen Interviews Trevor Traina about Kresus Labs
Stability AI’s Founder Made Misleading Claims — Report
Robert Scoble and Tim Draper
Editorial
Last week I took a risk regarding Apple’s face mask. To quote me:
If it comes, I will probably buy one (I just have to know what it does 😉 ), but I suspect it will soon sit idly on a shelf because there is nothing I want to use it for. Gaming is the only real use case, single and multi-user. And as it is mixed reality, it can extend to the real world, placing images and action there. Meta’s new device promises the same. And then there are enterprise uses.
Two things. I said Apple was too smart to get into this space. I was clearly wrong about Apple getting into the space. I hope I am not wrong about them being smart. No matter how good it is, how many people want to wear a face mask?
But I am asking myself about the statement: “it will sit idly on the shelf.” I will certainly buy one after watching the keynote and reading much of the commentary from @om, @gruber, and others. I would love to say that I will use it. For that to happen, a few things will need to be true:
If it can replace every screen in the house and clear my desk of everything. if it can replace my living room short-throw projector and all my AppleTVs. if it can replace my Mac and my Macbook Air. if it can replace my music system. If I can watch live sports on it, sitting in the stadium remotely with a 3d view of a game. Then It can come off the shelf.
It won’t (yet) replace my camera. It will not (yet) replace my iPhone or my Apple Watch because I would not wear it outside.
But just the things it could replace amount to many tens of thousands of dollars of equipment. And it may feel liberating to return to a minimalist, relatively clutter-free home. $3500 will not seem expensive.
As for the office? Whatever trend there is towards remote and hybrid working will be seriously accelerated if it is adopted. Facetime and document collaboration are easily imagined.
If it does all of those things, I will use it inside for work and leisure. It will not sit on the shelf. But… for that to be true, it will need to function through the day – probably for 10-15 hours – during work and evening leisure time. That will require a level of comfort and invisibility so my mind can function free of irritants. Battery life and fit will be key. And lenses that support my flaky eyes.
Honestly, I am somewhat optimistic that it could work. So yes, I think I was likely wrong. We will only know a year or so from now.
What would make me right? Humans do not adopt it because it is not feasible to wear it for long stretches. That remains a very strong possibility.
This week’s reading is filled with many takes on that and on the other big news of the week – The SEC charging Binance and Coinbase with various crimes. Sequoia Capital splitting into three firms based on geography, Marc Andreessen’s 7000-word essay on AI and responses, and Lionel Messi joining Inter Miami, a US soccer franchise backed by Apple’s cash. Enjoy.
Essays of the Week
First Impressions of Vision Pro and VisionOS
Wednesday, 7 June 2023
I got to spend about 30 minutes Monday afternoon using a Vision Pro and VisionOS at Apple Park, in a temporary “field house” building Apple constructed specifically for this experience. This was a one-on-one guided tour with someone from the Vision product marketing team, with the ground rules that I’m allowed to talk about everything I experienced, but photos and videos were not permitted.
It was a very fast 30 minutes, and the experience was, in a word, immersive. I’d pay good money just to run through the exact same 30 minutes again.
It was nowhere near enough time, nor was I able to wander far off the rails of the prepared demos. It’s very clear that the OS and apps are far from finished. But even given the brevity of the demo and constraints of the current state of the software, there are a few things I feel confident about describing.
First: the overall technology is extraordinary, and far better than I expected. And like my friend and Dithering co-host Ben Thompson, my expectations were high. Apple exceeded them. Vision Pro and VisionOS feel like they’ve been pulled forward in time from the future. I haven’t had that feeling about a new product since the original iPhone in 2007. There are several aspects of the experience that felt impossible.
Is it a compelling product, though? It’s a famous Steve Jobs axiom that technology is not enough, that you don’t make compelling products — let alone entire platforms — starting from advanced technology and working backward. You start with a vision for the product and platform experience and then create new technology to make it real. I simply can’t say whether Vision Pro is going to be a compelling product. I spent too little time with it, the software as of today is too far from complete, and, most importantly, the whole experience is too entirely new and mind-bending to render any such conclusion.
But the potential for Vision Pro to be a compelling product, across several use cases, is obvious. This might be great. And without question it is interesting, and I think the fundamental conceptual bones Apple has designed for VisionOS lay the groundwork for a long future. The first generation Vision Pro may or may not be a successful product — I simply don’t want to speculate on that front yet. But even just a small taste of VisionOS made me feel confident that it is going to be the next major platform for Apple and Apple developers, alongside MacOS and iOS/iPadOS.1
Apple’s Vision Pro Isn’t the Future
The new mixed-reality headset is an alarming misfire. Has Apple lost its innovation mojo?
I’M NOT A gambler, but I’d bet everything that Apple’s Vision Pro will flop.
When the $3,499 mixed-reality headset goes on sale in 2024, no doubt diehard Apple enthusiasts and VR/AR hobbyists will bring their sleeping bags and line up outside the Apple Store doors, hooting and hollering and having a ball. Maybe some gamers will get on board.
Do People Actually Want to Wear a Headset All the Time?
But the rest of us? No. Absolutely not. Don’t be ridiculous. This is not a “revolutionary” gadget, no matter how confident Tim Cook looks when he says it is. It’s a rare misfire, and a sign that Apple is losing its ability to turn tech-geek novelties into normie must-haves. It doesn’t augur the future so much as suggest that Cupertino doesn’t have a clear view forward.
“Every successful Apple product of the past two decades has disappeared into our lives in some way—the iPhone into our pockets, the iPad into our purses, the Apple Watch living on our wrists, and the AirPods resting in our ears,” my colleague Lauren Goode wrote this week, after demoing the device at WWDC. “But the Vision Pro is also unlike almost every other modern Apple product in one crucial way: It doesn’t disappear.” Instead, Goode wrote, the device settles onto your face, hides your eyes, “sensory organs that are a crucial part of the lived human experience.” The same was true of all virtual reality headsets and augmented reality glasses, she conceded, but the Vision Pro marked the first time an Apple product had made such an intrusion into people’s lives.
Reading Lauren’s review converted me into a full-fledged Vision Pro doomer. It drives home the reality that an Apple headset, no matter how nifty its specs, is still a big honking gizmo plonked between its wearer and the rest of the world, inherently a barrier more than a conduit.
Although Apple is positioning it as disruptive, Vision Pro is the latest in a long string of high-profile, splashy headsets designed to bring augmented reality, virtual reality, or both to the masses. Its predecessors include Meta’s Quest Pro, Microsoft’s HoloLens, the Magic Leap 2, and Google Glass—four mega-hyped headsets pitched as paradigm-shifting, all of which whiffed. Vision Pro is the newest, most advanced iteration of the idea, no doubt. That doesn’t change the challenge it faces, which is the same challenge that has hobbled every VR and AR headset that came before it: the whole alternative-reality headset thing.
Apple TV gets FaceTime, a new control center, findable remote and more
Sarah Perez@sarahintampa / 11:35 AM PDT•June 5, 2023
Image Credits: Apple
If you were playing WWDC23 bingo, you’d be missing a key space — the one you could mark if Apple again ignored Apple TV in its keynote address. This year, as it turns out, the company did deliver a small set of improvements to tvOS and its Apple TV device, though nothing as big as new hardware. With tvOS 17, Apple is redesigning the control center, bringing the FaceTime app to Apple TV (sorta!), and even helping you find a lost TV remote, among other things.
The new control center will help Apple TV owners do things like access their AirPods settings, home cameras, system controls, and user profiles with a touch of the remote.
You’ll also be able to now select favorite photos as your screensaver featuring “fluid animations and transitions” — a baseline feature other streaming devices, like Roku, already offer.
The bigger update in terms of family connection, however, is the introduction of FaceTime to Apple TV. The app will take advantage of Continuity Camera, which means the Apple TV will connect to your iPhone or iPad to use the camera and microphone, while the call itself is then shown on the big screen.
To use FaceTime, you’ll place your call on your iPhone and then move it to Apple TV, or you can pick up a call from a friend on your iPhone, then move it to Apple TV. In other words, the app is a companion to the FaceTime app you already use, rather than a standalone, native app experience you can use on the Apple TV on its own. However, by leveraging the camera on the iPhone or iPad, it can take advantage of features like Center Stage, which keeps you in the frame as you move around.
Plus, Apple says you can use SharePlay to watch TV with friends while on FaceTime, where everyone stays in sync.
Third-party apps, like WebEx and Zoom, will also be able to access a Continuity Camera API, Apple says. They’ll release updated apps by the end of the year with similar functionality, which will make things more interesting for the device from a business perspective.
Ultimate Guide to Platforms
Patterns and Practices in the Creation, Rise, and Fall of Platforms
JUN 1, 2023
The brass ring of product development is to create a platform. I say brass ring because by and large creating a platform is elusive for most products, at least at the scale we tend to think of when we think of platforms (Windows, iPhone, AdWords, and so on). In reality when, we look around there are tons of examples of platforms that deliver massive economics and powerful gains for those that bet on the platforms as customers and third parties.
While we all want to create a technology platform, we lack a common taxonomy or definitions. Once we have that definition, we can also look at the patterns that lead to the rise and fall of platforms.
The organization of this essay is as follows:
What is a Technology Platform?
A note on economics…
Models of Platforms
A note on stickiness/switching costs…
Why do platforms fail at launch?
A note on failing to consider all the 4 Ps…
Why do platforms succeed?
A note on platform moats…
The Platform Lifecycle
What is a Technology Platform?
The simplest definition of a platform is a technology that serves as an ingredient for the success of others but is so important that it achieves recognition and a market position on its own. The market position is such that there are many attempting to offer the same complete product or service, but the platform defines the not-so-secret ingredient upon which most, or even all, of the end-products or services rely.
Limiting the discussion to relatively modern business, one of the earliest examples of a platform strategy was undertaken at General Motors (GM) with the advent of automobile product lines. Within GM in the 1950s and later, what started out as efficiency in supply chain procurement turned into a go-to market strategy. By relying on common assemblies and other subsystems of cars while differentiating on branding and option choices, GM created the illusion of serving many more customers with cars uniquely tuned to specific customer segments. This eventually led to separate businesses based on a common platform, for example the Camaro, Firebird, and Trans-Am were very different customers for mostly the same car. An eye-opening list of just how far this platform strategy went can be found on the list of GM platforms.
Deliberately platformizing (my word) an existing product and attempting to reuse it over and over again is something that incumbents do once they achieve success. OREO cookies or Hershey chocolate appear in a dozen different aisles in supermarkets. Clorox bleach-based products fill an entire aisle. In toys and entertainment, from Barbie to Star Wars we have seen these platforms expand. Thus, the Disney Corporation today is one of the great executors of a platform strategy.
Historically in software, Windows attempted this strategy with a notion of the Win32 API serving as the basis for the strategy. As did Java, with the virtual machine at the core of an expansive platform. Both of these serve as cautionary tales when it comes to a deliberate expansion of a platform strategy.
Sequoia spin-offs show US geopolitical games are liability for Silicon Valley venture capital
By Hu Weijia Published: Jun 08, 2023 08:49 PM
Illustration: Chen Xia/Global Times
Sequoia Capital, one of Silicon Valley’s most prominent venture capital firms, is breaking itself up, separating its US and China operations by spinning out its Chinese unit into an independent company, the New York Times reported on Wednesday. The company said its global footprint had become “increasingly complex” to manage, but it is generally believed there are not only economic considerations, but also geopolitical elements behind Sequoia’s split.
Several days ago, a US Treasury official was quoted by the South China Morning Post (SCMP) as saying that Washington is working on new regulations that would limit capital flows into sensitive technologies in China such as semiconductor and artificial intelligence (AI). If the SCMP’s report is true, it is very likely that US-based venture capital firms such as Sequoia will be badly affected. Since it entered China in 2005, Sequoia has made bets on fast-growing Chinese tech companies such as Alibaba and TikTok parent ByteDance.
If Washington continues politicizing and weaponizing sci-tech issues, restricting the flow of US investment into Chinese companies working on chip, AI and quantum computing, US investors are likely to be the first and major victims. Sequoia’s dramatic move to split its business is, to some extent, a signal that the risk of escalating geopolitical tensions is a growing liability in Silicon Valley. Washington should take responsibility as the chief antagonist behind the increasing uncertainty.
In recent weeks, the US has frequently expressed its desire for dialogue with China. However, it seems there are questions within even domestically in the US about whether Washington genuinely intends to put an end to its high-stake chip war and stabilize economic relations with China. Sequoia’s decision reflects US business community’s concern about deteriorating relations between China and the US. This should sound the alarm: the US has witnessed increasing divergences between political and business communities over China-related issues. The Biden administration needs to correct its policy toward China, through its practical actions to stabilize bilateral ties and regain trust from the US business community.
According to the New York Times report, Sequoia has more than $53 billion in assets under management in the US and Europe, $56 billion in China and $9 billion in India and Southeast Asia. No matter how hard Washington works on its “decoupling” from China push, from the perspective of the US venture capital titan, the Chinese market remains irreplaceable.
The situation has become very clear now. Washington’s “decoupling” campaign can no longer continue. The US business community has moved to pressure Washington to stop playing games. An increasing number of executives of financial giants from the US and other Western countries have recently visited China and expressed their confidence and commitment to the Chinese market.
Video of the Week
AI of the Week
Why AI Will Save the World
TABLE OF CONTENTS
AI can make everything we care about better
The Baptists and Bootleggers of AI
AI Risk #1: Will AI kill us all?
AI Risk #2: Will AI ruin our society?
AI Risk #3: Will AI take all our jobs?
AI Risk #4: Will AI lead to crippling inequality?
AI Risk #5: Will AI lead to people doing bad things?
The actual risk of not pursuing AI
The era of Artificial Intelligence is here, and boy are people freaking out.
Fortunately, I am here to bring the good news: AI will not destroy the world, and in fact may save it.
First, a short description of what AI is: The application of mathematics and software code to teach computers how to understand, synthesize, and generate knowledge in ways similar to how people do it. AI is a computer program like any other – it runs, takes input, processes, and generates output. AI’s output is useful across a wide range of fields, ranging from coding to medicine to law to the creative arts. It is owned by people and controlled by people, like any other technology.
A shorter description of what AI isn’t: Killer software and robots that will spring to life and decide to murder the human race or otherwise ruin everything, like you see in the movies.
An even shorter description of what AI could be: A way to make everything we care about better.
Why AI Can Make Everything We Care About Better
The most validated core conclusion of social science across many decades and thousands of studies is that humanintelligence makes a very broad range of life outcomes better. Smarter people have better outcomes in almost every domain of activity: academic achievement, job performance, occupational status, income, creativity, physical health, longevity, learning new skills, managing complex tasks, leadership, entrepreneurial success, conflict resolution, reading comprehension, financial decision making, understanding others’ perspectives, creative arts, parenting outcomes, and life satisfaction.
Further, human intelligence is the lever that we have used for millennia to create the world we live in today: science, technology, math, physics, chemistry, medicine, energy, construction, transportation, communication, art, music, culture, philosophy, ethics, morality. Without the application of intelligence on all these domains, we would all still be living in mud huts, scratching out a meager existence of subsistence farming. Instead we have used our intelligence to raise our standard of living on the order of 10,000X over the last 4,000 years.
What AI offers us is the opportunity to profoundly augment human intelligence to make all of these outcomes of intelligence – and many others, from the creation of new medicines to ways to solve climate change to technologies to reach the stars – much, much better from here.
AI augmentation of human intelligence has already started – AI is already around us in the form of computer control systems of many kinds, is now rapidly escalating with AI Large Language Models like ChatGPT, and will accelerate very quickly from here – if we let it.
In our new era of AI:
Every child will have an AI tutor that is infinitely patient, infinitely compassionate, infinitely knowledgeable, infinitely helpful. The AI tutor will be by each child’s side every step of their development, helping them maximize their potential with the machine version of infinite love.
Every person will have an AI assistant/coach/mentor/trainer/advisor/therapist that is infinitely patient, infinitely compassionate, infinitely knowledgeable, and infinitely helpful. The AI assistant will be present through all of life’s opportunities and challenges, maximizing every person’s outcomes.
Every scientist will have an AI assistant/collaborator/partner that will greatly expand their scope of scientific research and achievement. Every artist, every engineer, every businessperson, every doctor, every caregiver will have the same in their worlds.
Every leader of people – CEO, government official, nonprofit president, athletic coach, teacher – will have the same. The magnification effects of better decisions by leaders across the people they lead are enormous, so this intelligence augmentation may be the most important of all.
Productivity growth throughout the economy will accelerate dramatically, driving economic growth, creation of new industries, creation of new jobs, and wage growth, and resulting in a new era of heightened material prosperity across the planet.
Scientific breakthroughs and new technologies and medicines will dramatically expand, as AI helps us further decode the laws of nature and harvest them for our benefit.
The creative arts will enter a golden age, as AI-augmented artists, musicians, writers, and filmmakers gain the ability to realize their visions far faster and at greater scale than ever before.
I even think AI is going to improve warfare, when it has to happen, by reducing wartime death rates dramatically. Every war is characterized by terrible decisions made under intense pressure and with sharply limited information by very limited human leaders. Now, military commanders and political leaders will have AI advisors that will help them make much better strategic and tactical decisions, minimizing risk, error, and unnecessary bloodshed.
In short, anything that people do with their natural intelligence today can be done much better with AI, and we will be able to take on new challenges that have been impossible to tackle without AI, from curing all diseases to achieving interstellar travel.
And this isn’t just about intelligence! Perhaps the most underestimated quality of AI is how humanizing it can be. AI art gives people who otherwise lack technical skills the freedom to create and share their artistic ideas. Talking to an empathetic AI friend really does improve their ability to handle adversity. And AI medical chatbots are already more empathetic than their human counterparts. Rather than making the world harsher and more mechanistic, infinitely patient and sympathetic AI will make the world warmer and nicer.
The stakes here are high. The opportunities are profound. AI is quite possibly the most important – and best – thing our civilization has ever created, certainly on par with electricity and microchips, and probably beyond those.
The development and proliferation of AI – far from a risk that we should fear – is a moral obligation that we have to ourselves, to our children, and to our future.
We should be living in a much better world with AI, and now we can.
AI is a tool
Published by MartinVarsavsky.net
Several well known thinkers have expressed concerns about the dangers of Artificial Intelligence (AI). Elon Musk said that AI could become an “existential risk for human civilization,” while the late physicist Stephen Hawking warned that it could “spell the end of the human race.” I acknowledge these concerns but to some extent; AI can indeed pose a tremendous risk, much like nuclear weapons. However, it’s important to emphasize that, akin to nuclear weapons, AI cannot initiate an attack on its own. It must be programmed to do so.
AI, at its core, is a tool, a mirror. It possesses no personal agenda, no inherent morality, and no autonomous motivation to cause harm or good. It merely reflects the intentions of its creators and users. The real danger isn’t the mirror itself, but the reflection it casts – a reflection influenced by human actions.
Deepfakes, sophisticated cyber-attacks, manipulative social engineering, autonomous weapons, and perpetuation of biases are just a few examples of how AI can be misused when reflecting malevolent human intentions. Each of these scenarios underscores that the true threat isn’t AI developing its own harmful agency, but rather, those who may use AI as a tool for harm. AI can cause enormous harm, but not on its own.
It’s also equally important to remember that the reflection can be one of great good. AI can be used to create incredible benefits for society when utilized responsibly and ethically. Here are some examples:
Healthcare Advancements: AI can improve patient care, enhance early disease detection, and aid in the development of new treatments and medicines. Combine with pharmacogenomics AI can make medical care more personalized and efficient.
Climate Change Mitigation: AI can help us analyze and understand climate patterns, predict natural disasters, and optimize renewable energy sources. It’s a great tool in our fight against environmental damage.
Education Personalization: AI can tailor educational content to individual students, helping them learn more effectively and at their own pace. It’s transforming the educational landscape, making learning more accessible and personalized.
Poverty Alleviation: By aiding in resource allocation, predicting economic trends, and providing access to digital services, AI can contribute significantly to poverty reduction efforts.
Space Exploration: AI can analyze vast amounts of data from space, help navigate rovers on distant planets, and even aid in the search for extraterrestrial life. It’s an essential tool in our quest to explore the universe.
Like any tool, the benefits or harms of AI largely depend on the hands that wield it. AI itself isn’t the risk; it’s the potential for misuse by those with malevolent intent. Our focus should be on fostering responsible AI development and usage, guided by strong ethical principles and societal norms. AI can augment evil but use properly it can augment kindness. And that should be our objective.
Follow Martin Varsavsky on Twitter: twitter.com/martinvars
Marc Andreessen Is (Mostly) Wrong This Time
Silicon Valley’s preeminent venture capitalist tries to craft the ur-narrative for generative AI, and in doing so lays bare its contradictions.
MARC ANDREESSEN OCCASIONALLY sets the world on its ear with a sweeping hypothesis about the dawn of a new technological era. In his legendary 2011 blog post “Why Software Is Eating the World,” the cofounder of Andreessen Horowitz made the then-novel, now-undeniable case that even the most old-school industrial companies would soon have to put software at their core. In 2020, as Covid-19 caught the world desperately short of masks and nasal swabs, he published “It’s Time To Build,” a call to arms for reviving investment in technologies that could solve urgent problems like pandemics, climate change, crumbling infrastructure, and housing shortages.
Now he’s back with a 7,000-word screed, another stab at framing the narrative; this time, the story is that “AI will not destroy the world, and in fact may save it.” Much of it is devoted to debunking AI doom scenarios, and the rest to touting AI as little short of a civilizational savior.
This is of course predictable. Andreessen invests in technological revolutions, so he has little incentive to do anything but hype them up. His post does have value, though, in two ways. First, its obvious blind spots are a useful guide to the thinking of the biggest AI hypesters and where they go astray. Second, its takedown of some of the more hysterical AI fears is actually (somewhat) on target.
So let’s dive in.
What AI Is, What AI Could Be
Andreessen tips his hand early by offering “a brief description of AI”: “The application of mathematics and software code to teach computers how to understand, synthesize, and generate knowledge in ways similar to how people do it” (my emphasis).
This seemingly innocuous parallel with human thinking, much like the phrase “artificial intelligence” itself, elides the vast gulf in capability between human minds and the current state of machine learning. Large language models (LLMs) are statistical inference algorithms. They predict the next likeliest thing in a sequence of things, such as words in a sentence. They produce what looks very much like human writing because they’ve been trained on vast quantities of human writing to predict what a human would write.
You’ll have already noticed that this is not even remotely similar to how you “understand, synthesize and generate knowledge.” You, like every human, have learned about the world by directly interacting with it. You’ve developed conceptions of physical objects such as trees and tables, of abstractions such as poverty and ethics, and of other people’s thoughts and feelings. You’ve learned to use language to talk about and process those conceptions, but language is just a layer for you, a way to share and refine your mental picture of the world. For LLMs, there is no mental picture; language is all there is.
To be sure, LLMs have made surprising leaps in ability recently, leading Microsoft researchers to claim that GPT-4, the latest model from OpenAI, contains “sparks” of general intelligence. And LLMs are not the only avenue of AI research. It can’t be ruled out that machines will eventually develop something more like our intelligence—though there are also good reasons to think it will end up being more alien than human.
OpenAI’s plans according to Sam Altman
Raza Habib
OpenAI’s plans according to Sam Altman
Last week I had the privilege to sit down with Sam Altman and 20 other developers to discuss OpenAI’s APIs and their product plans. Sam was remarkably open. The discussion touched on practical developer issues as well as bigger-picture questions related to OpenAI’s mission and the societal impact of AI. Here are the key takeaways:
1. OpenAI is heavily GPU limited at present
A common theme that came up throughout the discussion was that currently OpenAI is extremely GPU-limited and this is delaying a lot of their short-term plans. The biggest customer complaint was about the reliability and speed of the API. Sam acknowledged their concern and explained that most of the issue was a result of GPU shortages.
The longer 32k context can’t yet be rolled out to more people. OpenAI haven’t overcome the O(n^2) scaling of attention and so whilst it seemed plausible they would have 100k – 1M token context windows soon (this year) anything bigger would require a research breakthrough.
The finetuning API is also currently bottlenecked by GPU availability. They don’t yet use efficient finetuning methods like Adapters or LoRa and so finetuning is very compute-intensive to run and manage. Better support for finetuning will come in the future. They may even host a marketplace of community contributed models.
Dedicated capacity offering is limited by GPU availability. OpenAI also offers dedicated capacity, which provides customers with a private copy of the model. To access this service, customers must be willing to commit to a $100k spend upfront.
2. OpenAI’s near-term roadmap
Sam shared what he saw as OpenAI’s provisional near-term roadmap for the API.
2023:
Cheaper and faster GPT-4 — This is their top priority. In general, OpenAI’s aim is to drive “the cost of intelligence” down as far as possible and so they will work hard to continue to reduce the cost of the APIs over time.
Longer context windows — Context windows as high as 1 million tokens are plausible in the near future.
Finetuning API — The finetuning API will be extended to the latest models but the exact form for this will be shaped by what developers indicate they really want.
A stateful API — When you call the chat API today, you have to repeatedly pass through the same conversation history and pay for the same tokens again and again. In the future there will be a version of the API that remembers the conversation history.
2024:
Multimodality — This was demoed as part of the GPT-4 release but can’t be extended to everyone until after more GPUs come online.
3. Plugins “don’t have PMF” and are probably not coming to the API anytime soon
A lot of developers are interested in getting access to ChatGPT plugins via the API but Sam said he didn’t think they’d be released any time soon. The usage of plugins, other than browsing, suggests that they don’t have PMF yet. He suggested that a lot of people thought they wanted their apps to be inside ChatGPT but what they really wanted was ChatGPT in their apps.
New Statement Signed by the Likes of OpenAI’s Sam Altman Warns of AI’s Extinction Risk
A new statement by the Center for AI Safety, a San Francisco-based not-for-profit, warns of the existential risks associated with AI. Some who signed on to this statement include OpenAI’s Sam Altman, DeepMind CEO Demis Hassabis, MIT’s Max Tegmark, Microsoft CTO Kevin Scott, and many other notable names.
In part, the statement points to the risk associated with AI and equates the harm to that of a nuclear apocalypse. So the statement calls on policymakers to focus on AI, mitigating any risks that could harm humans. This isn’t the first time scientists and other AI leaders have gone public with their concerns about AI, but this is the first time such strong language was used.
Back in March, 1,000 tech leaders, scientists, and thought leaders called for a pause in the development of large language models more powerful than GPT-4. Their concerns centered around the unforeseen impacts of the technology before society was ready to handle it.
The statement from the Center of AI Safety used much stronger language and said in part, “Mitigating the risk of extinction from AI should be a global priority alongside other societal-scale risks such as pandemics and nuclear war.”
These words are pretty strong, and will for most, invoke some of sci-fi’s most popular tropes such as AI gone wrong, and killer robots. But it seems that the Center for AI Safety has this in mind and is worried that “AI’s most severe risks” could find themselves lost in the noise related to AI.
To the group, AI the risks associated with AI require an “open…discussion.” For Sam Altman and OpenAI’s part, they announced on Twitter new grants aimed at helping AI become “democratic” through ten $100,000 grants.
Overall, it seems that OpenAI’s Sam Altman and other signatories are not alone. Though slowly, multiple nations, including the United States, China, and members of the EU are taking action when it comes to AI regulation. Each are at different stages of development and seems to take aim at different issues ranging from privacy, fraud, and more.
It’s clear that nations are taking reports, such as Goldman Sachs AI report seriously, as AI continues to enter multiple industries at a rapid pace. As of right now, the calls for policymakers to take on AI regulation continue to grow.
News Of the Week
New Unicorns In May Hit Double Digits With Two AI Companies
June 8, 2023
Ten companies joined The Crunchbase Unicorn Board in May 2023 — double the count for April 2023 but still significantly down from the 34 new unicorns in May 2022.
This new unicorn count took place in a funding environment where the most active unicorn investor, Tiger Global, is looking to sell its stakes in private companies, and investors continue to downgrade unicorn portfolio values.
The 10 companies hailed across nine different industries, and AI topped the list with two companies. Other sectors with a single new unicorn ranged from energy, biotech, robotics, enterprise SaaS and transportation among others.
Five of the new unicorns are U.S.-based and two are from China. Canada, Indonesia and Japan each counted one new unicorn this past month.
Around $5 billion of the $22 billion in global venture funding raised in May 2023 went to unicorn companies. That’s less than half of the $11 billion that unicorns raised in May 2022. The largest funding last month went to fast-fashion e-commerce platform Shein, which raised $2 billion in a single round at a reduced valuation of $66 billion.
Here are the new unicorns:
AI
Toronto-based Cohere, a generative AI large language model developer for enterprises, raised $270 million in its Series C funding. The funding was led by Inovia Capital valuing the 4-year-old company at $2.2 billion.
Generative video AI company Runway, based out of New York, raised a $100 million Series D led by Google. The funding valued the 5-year-old company at $1.5 billion.
Energy
China-based Ryefield Energy, a battery energy storage company, raised $140 million led by Sparkedge Capital. The company was valued at $1.4 billion.
Agtech
Indonesia-based eFishery, developer of an IoT feeding device for shrimp and fish, and an e-commerce platform for aquafarmers, raised a $108 million Series D led by Abu Dhabi PE fund G42 Expansion Fund. The company was valued at $1.3 billion.
Cleantech
Massachusetts-based Gradiant helps enterprises dependent on water reduce and treat wastewater. It raised a $225 million Series D led by BoltRock Holdings and Centaurus Capital in a round which valued the company at $1 billion.
Enterprise SaaS
Restaurant management software company Restaurant365 based in Irvine, California, raised a $135 million round which valued the company at $1 billion. The funding was led by private equity firms Kohlberg Kravis Roberts and L Catterton.
Proptech
Proptech company Avenue One provides a platform of services from brokers, contractors and property managers for buyers of rental properties. The company based in New York raised $100 million led by WestCap valuing the company at $1 billion.
Transportation
Tokyo-based GO, the leading ride hailing app in Japan, raised a $72 million Series D led by Goldman Sachs which valued the company at $1 billion. The company is also developing software to monitor safe driving practices.
Robotics
Humanoid robot developer Zhiyuan Robotics based in Shanghai, raised its third funding led by Baidu Capital, valuing the company at $1 billion. The amount raised was not disclosed.
Biotech
VectorBuilder, a gene delivery technology company headquartered in Chicago, raised a funding which valued the company at $1 billion.
Mercedes Just Beat Tesla to This Huge Milestone in the Race to Self-Driving Cars
Mercedes drivers in California may soon be able to scroll TikTok while their car does all the driving for them.
3 HOURS AGO
An offshoot of the AI race revolves around self-driving cars.
And in that realm, despite Tesla’s gains with its Full-Self Driving (FSD) technology, Mercedes just beat out the competition.
The California Department of Motor Vehicles approved Mercedes-Benz to sell or lease self-driving cars to the public June 8.
Similar to Tesla’s Autopilot, Mercedes’ Drive Pilot allows for hands-free driving. But it goes a bit beyond simply hands-free, allowing the driver to turn away from the road in something that is essentially eyes-free.
Drivers do, however, have to be prepared to take control if necessary, so no naps. If the car’s internal cameras can’t see the driver’s face, the system will disengage.
Drive Pilot, according to Mercedes, will only be enabled at speeds below 40 mph on “suitable freeway sections” during daylight hours.
Mercedes is the first carmaker to be approved to sell or lease self-driving cars in California.
GM EV owners will gain access to Tesla Superchargers in 2024
General Motors EVs will adopt the NACS standard in 2025
hapabapa via Getty Images
Sean Buckley | @seaniccus | June 8, 2023 8:21 PM
Ford isn’t the only electric automaker switching to Tesla’s North American Charging Standard — General Motors says it’s making the change, too. CEO Mary Barra announced the move during a Twitter Spaces chat with Tesla CEO Elon Musk on Thursday, stating that its electric vehicles will make the NACS open-source connector standard in all GM EVs in 2025.
As part of the collaboration, all GM EVs will gain access to 12,000 Tesla Superchargers in 2024. Drivers of existing GM EVs won’t have to upgrade their vehicles to use Tesla’s chargers, but will need to use an adapter to make their vehicle compatible. Likewise, GM says it will be developing an adapter that will allow future NACS-enabled EVs to charge its existing network of CSS-capable fast charging stations.
That backwards compatible charger could prove important. Tesla’s willingness to open up its charging system to non-Tesla vehicles was originally announced alongside a $7.5 billion Biden administration initiative to expand EV charger availability in the US — but that plan heavily indexes on building out CCS chargers.
“This collaboration is a key part of our strategy and an important next step in quickly expanding access to fast chargers for our customers,” Barra said in GM’s statement on the partnership. “Our vision of the all-electric future means producing millions of world-class EVs across categories and price points, while creating an ecosystem that will accelerate mass EV adoption.”
Sequoia Splits Its Global Fund Into 3 Separate Businesses
June 6, 2023
Sequoia Capital announced this morning that it is breaking its global fund into three independent businesses that will chart their own paths going forward.
Sequoia will continue to invest in the U.S. and Europe. Its China business will keep its name in Chinese and be known as HongShan in English. And the India and Southeast Asia business will become Peak XV Partners.
“It has become increasingly complex to run a decentralized global investment business,” the firm said in its announcement signed by the leaders of each of its regional funds. But more significantly, as local companies compete on a global basis, the firm is seeing more portfolio conflicts which could lead to passing on opportunities to invest in competitive sectors.
Even within a region, the firm faces an issue with competition. Sequoia divested from Finixdue to perceived competition in payments with its portfolio company Stripe. In an interview with Forbes, the firm said the breakup is not due to geopolitical tensions.
‘Patient capital’
Sequoia Capital had already made a significant change to its fund structure in recent years. It announced that a single fund for its U.S. and European business would distribute to sub-funds, stating the “10-year fund cycle has become obsolete.” As part of that November 2021 announcement, months before the market correction, Sequoia restructured itself as a registered investment adviser. This new role allowed the firm to set itself up to retain stock when its private companies go public, and to invest in public companies as well as opportunities in crypto and Web3.
In an early 2021 interview, the firm’s steward Roelof Botha emphasized the firm’s “patient capital” approach. The firm will often wait years before it disperses shares to shareholders, he said.
“Sometimes people don’t realize how patient we are,” Botha said “And we’ve earned this right with our LPs, to have patience with distributions.”
The regional funds have operated from the start with a “local-first” approach. The India fund was launched in 2000 and the China-based fund in 2005. Each business will become fully independent by March 31, 2024.
Twitter’s Ad Revenue Continues to Decline Under Elon Musk
Published June 5, 2023
By Andrew Hutchinson – Content and Social Media Manager
While Elon Musk has continuously reiterated that his purchase of Twitter is about securing freedom of speech, and that recouping the $44 billion he paid for the company doesn’t matter, he still has bills to pay, and investors to make whole as he goes about reforming the business.
Which is why the latest reports on Twitter’s ad sales are important, with The New York Times reporting that Twitter’s US ad sales are down 59% year-over-year, reflecting worsening ad sales at the company over time, despite Musk’s assurances that advertisers had been steadily coming back to the platform.
According to NYT, Twitter recently shared an internal overview of its ad sales performance, which showed that total ad sales are down 59%, while the company is regularly falling short of its US weekly sales projections. That’s worse than the 50% decline that Musk confirmed back in March, which suggests that businesses are still spooked by Musk’s reformations at the app, which have included new rules around acceptable speech, and reinstatements of some of the platform’s most controversial users.
And Musk himself hasn’t helped its case. He continues to share his often controversial opinions on hot-button topics, including gender affirmation, the war in Ukraine, the government’s COVID response, population collapse, crime, immigration, and more.
Which, of course, was the whole point of his Twitter takeover in the first place. Musk says that he bought Twitter to fight for free speech, and battle against mass media censorship, and that, in some ways, is a noble stance. But the side effect is that Twitter’s ad revenue is going to take a hit as a result.
And when that income stream makes up some 90% of your company’s revenue, that is a challenging path to take.
Musk’s most recent stance on this front saw two of his top brand safety experts move on from the company, after Musk sought to reverse their decision on supporting the release of a controversial anti-trans documentary. Musk demanded that the rules around such be watered down, while also promoting the documentary on his own profile, which led to both Ella Irwin, the platform’s head of trust and safety, and AJ Brown, Twitter’s chief of brand safety, leaving the company as a result.
SEC Files 13 Charges Against Binance, Founder Changpeng Zhao And Related Entities
June 5, 2023
The Securities and Exchange Commission filed 13 charges against the world’s largest crypto exchange Binance, its co-founder Changpeng Zhao, and other entities related to the exchange.
In the complaint, the SEC alleges Zhao and Binance controlled customers’ assets on the platform and commingled or diverted billions of dollars of those assets as they desired, including to Sigma Chain, a European business Zhao owned and controlled.
It further alleges that while Zhao and Binance told U.S. customers they were restricted from transacting on the Binance.com international exchange, they allowed “high-value U.S. customers” to trade on the platform.
U.S. customers instead were told they could transact on what was supposedly the independent trading platform Binance.US, but in fact that exchange was actually controlled by Zhao and Binance, the complaint also alleges.
In a blog, Binance said it is disappointed by the SEC’s decision to file charges and had been cooperating with investigators.
In the news
Of course, this is not the first time Binance and its founder have popped up in the news, having played a role in the collapse of FTX.
Last fall, Changpeng Zhao tweeted that Binance would sell its holdings of FTX’s native token, FTT, due to “recent revelations that have came to light” about the token.
FTT prices dropped like a rock amid the controversy and FTX was thrown into disarray as withdrawals mounted.
Binance seemed to emerge as FTX’s knight in shining armor when Zhao tweeted the newshis company would acquire FTX, but that deal quickly fell apart.
That led to the complete collapse of FTX and its founder’s, Sam Bankman-Fried, empire.
Finally in December, the SEC filed charges against disgraced FTX founder Bankman-Fried for “orchestrating a scheme to defraud equity investors.”
Binance actually was one of FTX’s first lead investors, taking part in a round of undisclosed value in late 2019, according to Crunchbase data.
Andrew Keen and Trevor Traina discuss Kresus Labs
Stability AI’s Founder Made Misleading Claims — Report
June 5, 2023
Stability AI’s founder Emad Mostaque made exaggerated statements about both his own background and his generative AI startup, according to a Forbes report.
The report, published over the weekend, claims Mostaque misled people — including investors — about receiving a master’s degree from Oxford University, as well as a deal with Amazon — which referred to it as a strategic partner — but was nothing more than a standard cloud computing leasing deal.
Stability has also described the OECD, WHO and World Bank as partners to investors — something they all deny, per the report.
The report also detailed issues related to payroll uncertainty, unpaid invoices, and monetary transactions between Stability and accounts controlled by Mostaque and his wife. A Stability spokesperson said those were actually loans the couple made to and from the company.
Controversy and big money
This isn’t the first time the London-based startup has found itself in the headlines.
Last fall, Stability AI locked up a $101 million raise led by Coatue, Lightspeed Venture Partners and O’Shaughnessy Ventures. The company did not release a valuation at the time, but Bloomberg reported the new cash infusion valued the company at around $1 billion.
At the time, some AI researchers disputed the startup’s claims that it created the image generator Stable Diffusion — which was an open-source project developed by researchers.
It also was reported at that time the startup was looking to raise an additional $1 billion of capital at a multibillion-dollar valuation.
Those funding talks have stalled, per Forbes.
Stability is an AI-driven visual art startup. At the time of its fundraising, the company claimed its consumer-facing product DreamStudio had grown to more than a million registered users who have created more than 170 million images.
Startup of the Week
Apple’s Messi Vision
Which is decidedly not messy, and perhaps rather brilliant…
Well, I didn’t see Apple’s long cultivated ‘Services’ narrative going in thisdirection. Earlier today, it was announced that Lionel Messi, the world’s greatest soccer player,¹ would be leaving Europe behind and joining Inter Miami.² This matters because the biggest name in fútbol (soccer), and arguably in all sports, is coming to America.³ But it also matters because he’s coming to play in the MLS. A league that has an exclusive television deal with Apple. And as such, Apple apparently helped lure Messi over with an extraordinary offer: a cut of revenue from new subscribers to their Season Pass service.
That’s right. Apple is using the world’s highest-profile athlete as lead gen.
Aging soccer stars coming to the US is nothing new. Everyone from Pelé to David Beckham — who, not coincidentally now is an owner of Inter Miami — have come to play out the twilight of their careers, essentially as a marketing maneuver for both the leagues and the stars. The difference here is Apple. The nearly $3 trillion company. The one that makes Macs and iPads and iPhones and soon Vision Pros (more on that in a bit). They’re footing part of this bill with their Services revenue. Wild times.
And it actually makes a lot of sense on a few fronts.
First, Apple is obviously incentivized to help MLS grow in popularity given their aforementioned exclusive media deal. Simple.