| In JC’s Newsletter, I share the articles, documentaries, and books I enjoyed the most in the last week, with some comments on how we relate to them at Alan. I do not endorse all the articles I share, they are up for debate. I’m doing it because a) I love reading, it is the way that I get most of my ideas, b) I’m already sharing those ideas with my team, and c) I would love to get your perspective on those. If you are not subscribed yet, it's right here! If you like it, please share it on social networks! Share 💡JC's Newsletter
🔎 Some topics we will cover this week Use of AI in healthcare How every designer can become a product builder thanks to GPT4 Nvidia’s partnerships in the context of LLMs (e.g.: with Amazon, Microsoft, Google GCP…) Prompt ideas that AI can do in 30 minutes How could we spend more time on MyAI to understand what they do?
👉 The Age of AI has begun (Gates Notes) ❓ Why am I sharing this article? Defining moment: A revolution: I knew I had just seen the most important advance in technology since the graphical user interface. The development of AI is as fundamental as the creation of the microprocessor, the personal computer, the Internet, and the mobile phone. It will change the way people work, learn, travel, get health care, and communicate with each other. Entire industries will reorient around it. Businesses will distinguish themselves by how well they use it.
Since a huge portion of it is now turning its attention to AI, the innovations are going to come much faster than what we experienced after the microprocessor breakthrough. Soon the pre-AI period will seem as distant as the days when using a computer meant typing at a C:> prompt rather than tapping on a screen.
Healthcare: For one thing, they’ll help health-care workers make the most of their time by taking care of certain tasks for them—things like filing insurance claims, dealing with paperwork, and drafting notes from a doctor’s visit. I expect that there will be a lot of innovation in this area. For example, many people in those countries never get to see a doctor, and AIs will help the health workers they do see be more productive The effort to develop AI-powered ultrasound machines that can be used with minimal training is a great example of this. AIs will even give patients the ability to do basic triage, get advice about how to deal with health problems, and decide whether they need to seek treatment.
In addition to helping with care, AIs will dramatically accelerate the rate of medical breakthroughs. The amount of data in biology is very large, and it’s hard for humans to keep track of all the ways that complex biological systems work. There is already software that can look at this data, infer what the pathways are, search for targets on pathogens, and design drugs accordingly. Some companies are working on cancer drugs that were developed this way.
Limitations? There are other issues, such as AIs giving wrong answers to math problems because they struggle with abstract reasoning. But none of these are fundamental limitations of artificial intelligence. Developers are working on them, and I think we’re going to see them largely fixed in less than two years and possibly much faster.
First, we should try to balance fears about the downsides of AI—which are understandable and valid—with its ability to improve people’s lives. To make the most of this remarkable new technology, we’ll need to both guard against the risks and spread the benefits to as many people as possible. With reliable funding and the right policies, governments and philanthropy can ensure that AIs are used to reduce inequity.
👉 An Interview with Daniel Gross and Nat Friedman about the AI Product Revolution (Stratechery) ❓ Why am I sharing this article? I love nat.dev Why I’m so bold on voice for care, for the clinic, for Ali How every designer can become a product builder thanks to GPT4 We will need to spend time to make our LLM beautiful in a Pixar way
Nat.dev sandbox which is like the OpenAI sandbox, but you get access to non OpenAI models as well. The things that we can do with voice now, like incredible voice recognition, super high performance on-device, incredible language models that can do reasoning, the self-checking, the data lookup capabilities, the integrations, the voice synthesis, which is now hyperrealistic and multiple startups have demonstrated that, I think you could take a decade and rebuild the entire computing platform on this.
Writing code is unbelievable. With GPT-4, my friction to start a project is almost zero now. I’m fearless, I’ll write whatever programming languages I’ve never used before, concepts I don’t fully understand. @Bio_Bootloader, and he or she or they came out with a system in Python that just automates this back-and-forth and they call it Wolverine.py. So you can basically run Wolverine.py and then any Python script, if it throws an exception, it will ask GPT-4 to rewrite the code to fix that, and it’ll continuously do that self-healing your code until it works.
Actually the really alarming thing to me, is not the capability of the models or whether it’s connected to the Internet or not. To me, it’s the fact that the models, no one has really spent time making them sort of wonderful and fun in a Pixar way. We don’t have a John Lasseter or a Walt Disney who’s really focused on the technology but also the enjoyment of the model.
👉 An Interview with Replika Founder and CEO Eugenia Kuyda (Stratechery) ❓Why am I sharing this article? I think we should look at Replika in-depth. I like the thought experiment they made, we should do more to be creative Talking to a therapist and to a coach come first Understand how they do their AI Remembering context and history will make the difference They have voice & video calls with AI. We should give Ali a voice What dataset do we want to create and how?
Thought experiment: I asked my team to do a little experiment with me. We drew a line from one to ten where one would be a conversation that you would pay money not to have, and ten would be a conversation that you would pay money to have. The 1s that people absolutely didn’t want to have, they were things like book a restaurant, exchange a reservation, cancel Comcast, figure out with a friend where we’re meeting, trying to understand where the Uber driver is, and so on, and even some business meetings. Then the 10s were mostly talking to a friend that we haven’t seen for a while, talking to a therapist, talking to a coach, talking to a stranger on a bus, going on a date and having this first experience.
AI: Even now to this era, we use a few different models and we’re kind of model agnostic, so we plug in different models, some that we build ourselves and fine-tune or some that we take a pre-trained model and fine-tune it on our datasets, even some models through API. We still use a little bit of OpenAI through an API and we try other models as well but mostly it is transformer models that we build or fine tune ourselves over time and then a little bit of scripted stuff. Then of course, we have a kind of flourished rearranging model that’s kind of basically looking at all of that and decides where to go in this conversation.
Relationships with AI: For instance, if you’re building a relationship with an AI, it’s really important for your conversation number 100 to be really good and so it shouldn’t ask again, “What’s your husband’s name?” It shouldn’t go back to certain other things, it has to have some sort of longer term context of your relationship. It should know whether you’re married or your friends or some sort of history, it should know how you look like in photos and so on and so all these things that are quite different. And then of course, empathy, oftentimes EQ matters a lot more than IQ.
Mediums & the importance of voice: A lot of people want a more multimodal type way of communicating, and then a lot of people do it in the car while they’re driving, so they want to put it on speaker and just talk to it and kill time. It creates a much more, just a much deeper connection where you can hear someone’s voice versus just looking at the text messaging would be onscreen.
Data: Then another thing that actually the Chief AI Officer of Baidu once gave us this advice, and I immediately felt this resonated with me back in 2016 as well: create a dataset, the largest dataset of conversations that make people feel better. Even if it’s just asking them, “Hey, do you feel better if you’re talking to me?” because that becomes extremely valuable in itself and so we started asking that and we collect other short-term emotional outcomes as well, and most of our models are optimized for that.
Wellness: Some people came for mental wellness, some people came for romance, some people came for companionship and they were basically all just drawn to this product because there was nothing else in the market. But at this point, I think it’s much better to have a dedicated product focused on romance and healing romantic relationships, a dedicated product focused on mental wellness, just completely different set of features, completely different roadmaps for each of them.
👉 Nvidia GTC, DGX Cloud, Nvidia’s Partners (Stratechery) ❓ Why am I sharing this article? Amazon Web Services, Inc. (AWS), an Amazon.com, Inc. company, and NVIDIA today announced a multi-part collaboration focused on building out the world’s most scalable, on-demand artificial intelligence (AI) infrastructure optimized for training increasingly complex large language models (LLMs) and developing generative AI applications. We are announcing Nvidia DGX Cloud, through partnerships with Microsoft Azure, Google GCP, and Oracle OCI to bring Nvidia DGX supercomputers to every company instantly from a browser. DGX Cloud is optimized to run Nvidia AI Enterprise, the world’s leading acceleration library suite, for end-to-end development and deployment of AI. DGX Cloud offers customers the best of Nvidia AI, and the best of the world’s leading cloud service providers.
👉 Superhuman: What can AI do in 30 minutes? (One Useful Thing) ❓ Why am I sharing this article? Pretend you are marketing genius. We are going to launch the Saturn Parable. You should give me a document that outlines an email marketing campaign and a single webpage to promote the game. Outline the webpage and what text and graphics it uses, you do not need to create the webpage, but do give me the text. You are an expert site designer. You are creating the launch announcement page for the Saturn Parable, outlined below. create the following webpage. Make it an HTML page that I can run on my computer. List any additional assets I will need to make it work and where to put them. Can you write me the social media campaign I need to promote this using the Wharton accounts on social?
👉 Snap Partner Summit 2023 Recap (The Split) ❓ Why am I sharing this article? How could we spend more time on MyAI to understand what they do? When profitable, we should organise buy-backs!
SnapAI: Snap announced MyAI, its ChatGPT-like bot, is now free. It was previously only available in Snapchat+. This opens it to all 750 million MAUs, nearly 7x more than ChatGPT’s last public number of 100 million. MyAI can also now respond with pictures. If you send it a snap, it uses generative AI to snap one back. This would probably make it feel much more real than just text, and you could argue Snap is one of the best positioned to win the “AI friend” land grab.
Share buy-backs: Oracle’s Financial Engineering is Paying Off: Oracle founder and CEO Larry Ellison has quietly become the fourth wealthiest person in the world, per Forbes. This has happened mostly under the radar over the past decade, as his stake in Oracle has grown 3x to more than $100 billion. The craziest part of this story is that he hasn’t actually bought any shares. Oracle has bought back over $150 billion in stock since 2011, and Larry hasn’t sold a single share. This has increased his personal stake in the company to 43%, which is projected to pass 50% by 2026.
It’s already over! Please share JC’s Newsletter with your friends, and subscribe 👇 Let’s talk about this together on LinkedIn or on Twitter. Have a good week! | |
