10 Charts That Capture How the World Is Changing (Part II)
From AI to Healthcare, Remote Work to Black Friday Numbers
This is a weekly newsletter about how tech and culture intersect. To receive Digital Native in your inbox each week, subscribe here:
10 Charts That Capture How the World Is Changing (Part II)
In the last Digital Native, I shared 10 charts that I find interesting and emblematic of broader themes. The charts were wide-ranging, covering topics from homeownership to climate change, media consumption to job growth.
This week, I posed the same question to my partners at Index. I asked them to share a chart / trend that they’re paying attention to, and a few words on why. Below, I’m sharing a few of their responses. The topics covered here are again varied—healthcare, AI, remote work. Then I’ll wrap up this week with a few more charts of my own.
Let’s jump in 📈
Healthcare Administration Bloat | Molly Alter
My partner Molly Alter shared this chart on healthcare administration bloat:
As someone whose entire family works in healthcare (black sheep over here 🙋♂️), this chart resonates with me. In Molly’s words—
How many times have we filled out the same intake paperwork at the doctor’s office? Or sat on hold with an insurance company only to have them tell you that you’re not authorized for a procedure and you’ll have to pay out of pocket? Has anyone even seen their full health record? We can all feel what this chart displays: healthcare administration has gotten completely out of hand. Another stat which aligns with this trend is that at least 70% of healthcare providers still exchange data using a fax machine. Easing the administrative load on the healthcare system would allow hospitals to focus their energies on care delivery itself, and create far better experiences for patients themselves. If you’re building in this space, let me know!
This chart ties back to last week’s chart on price changes for goods and services. While smart TVs have dropped in price 98.5% since 2000, healthcare costs have swelled 100%. Much of Molly’s work at Index focuses on software that can bring down costs and broaden access in healthcare. You can reach her on Twitter at @molly_alter.
Remote Work Options | Hannah Seal
My partner Hannah Seal shared this chart on the availability of remote work. Fittingly, Hannah led our Seed and Series A investments in Remote, which builds on this trend by facilitating global HR and payroll for distributed teams.
As Hannah puts it:
The future of work is flexibility. According to McKinsey, 58% of job holders in the US work can work remotely for all or part of the week. A remote-first approach allows companies to hire the best person for the job, no matter where in the world they may be based. Talent is universal and companies are no longer restricted to a certain city or time zone, which creates huge opportunity for talent globally. However, the growth of remote and hybrid work also creates challenges around ways of working, communication, and coordination, and I’m excited to see the companies that will be built off the back of solving these problems.
In many ways, this chart ties back to last week’s chart on homeownership. For years, new homes were consistently located 10-15 miles from prior homes. This year, the median distance soared 3x+ to 50 miles.
This increase is pandemic-driven, with geography becoming less of a constraint on work. But as Hannah noted (you can find her on Twitter at @hannahlseal), remote work also begets challenges for organizations. We’ve seen teams double down on popular software tools that facilitate collaboration. At Index, we’ve invested in many of them: Notion, Figma, Dropbox, Slack. But there are also promising up-and-coming startups building for this new work reality: Tango, for instance, facilitates how-to guides that improve knowledge transfer; Guru is a company wiki for organizing company information; and Rewatch helps teams draw insights from video meetings.
The next iconic collaboration / productivity software companies will be purpose-built for a remote / hybrid world of work.
Periodic Table of Amazon Web Services | Kelly Toole
My partner Kelly Toole shared this visualization. Remember the periodic table? Here it is again (ignore those flashbacks to AP Chem), but depicting the bounty of services AWS offers:
Kelly (Twitter: @ckellytoole) writes:
Next March, it’ll be AWS’s 17th birthday. 17! We’re almost two decades into the Great Migration to the Cloud.
I recently read a blog post written by Modal founder Erik Bernhardsson. In it, he talks about how early we truly are in the evolution of this Cloud, despite the fact that it may feel like we’ve long been in it. One of his opening points is that AWS jumpstarted this paradigm shift with just a few services. Now, it has over 200 and counting. The image above helps to illustrate just how vast the surface area of AWS is today. It covers everything from gaming to robotics.
Yet despite this (and 20-odd years of development), AWS can still be incredibly hard to use! Even for simple things. I laughed out loud—in agreement—when I read one of Erik’s hopes for the future Cloud: “I never ever again want to think about IP rules. I want to tell the cloud to connect service A and B!”
We know that Java became hugely popular because it offered a new layer of abstraction above important programming concepts like memory management. By removing the need for engineers to deal with tedious (and bug prone) tasks like this, they could be more productive elsewhere. Folks still write plenty of C, but Java’s popularity demonstrates that well-designed layers of abstraction can truly be game-changing.
I suspect that we’ll see the same evolution happen here, where new generational companies will emerge that remove the need for certain categories of application developers to understand and write infrastructure code at all. Out-of-the-gate, those that focus on ease of use (and debuggability) will have an advantage over AWS and others.
Erik’s blog post was picked up byHacker News and sparked a long (192 comment) debate. Clearly, the developer community is keenly interested in seeing where these new Cloud abstraction(s) net out. As are we!
AWS is an incredible business. In The TikTokization of Everything, I shared this chart of AWS revenue growth (with 35% profit margins to boot):
But as Kelly shares, AWS is far from perfect. The arc of technology bends toward ease-of-use and broad accessibility, and we’re still in the early innings.
Midjourney v3 vs. Midjourney v4 | Cat Wu
In September’s When Art and Technology Collide, I wrote about Midjourney’s text-to-image generative AI. Midjourney has continued to up its game. My partner Cat Wu shared this visualization comparing Midjourney v3 and v4. Both sets of images are for the same prompt—“A penguin in Venice”—but Midjourney v3 is on the left and Midjourney v4 is on the right. It’s quite a leap forward.
Over the past year, the generative AI space has been moving at a breakneck pace. One area where we can clearly see this is in the progress of image models. Midjourney is an independent research lab that develops image models for a specific user persona: artists and creatives. Above, we see the dramatic improvement in Midjourney’s model over the last five months. On the left, we see the images produced by Midjourney Version 3 (released June 2022) which include a penguin with two beaks and another with three legs; on the right, we see the images produced from the same prompt by Midjourney Version 4 (released November 2022) which include fashionably-dressed characters and detailed backgrounds.
Looking forward, I expect these image models to continually improve in accuracy and resolution, making them more powerful tools for concept art, self-expression, and entertainment for both hobbyists and companies that need visual assets. I’m excited about giving more individuals the ability to bring their visions to life, and for this to unlock new visual concepts and bring more art into the world. (Note: We see comparable levels of improvements in large language models, which have far-reaching implications for every knowledge worker.)
Prior to joining Index, Cat worked at two Index-backed companies, Scale AI and Elementl. She’s the expert on all things AI, and you can find her on Twitter at @catherwu.
A lot of people worry about generative AI obfuscating human creativity. Rather, I see generative AI as a tool that will amplify creativity. I like how Midjourney’s founder, David Holz, puts it:
We don’t think it’s really about art or making deepfakes, but—how do we expand the imaginative powers of the human species? And what does that mean? What does it mean when computers are better at visual imagination than 99 percent of humans? That doesn’t mean we will stop imagining. Cars are faster than humans, but that doesn’t mean we stopped walking. When we’re moving huge amounts of stuff over huge distances, we need engines, whether that’s airplanes or boats or cars. And we see this technology as an engine for the imagination. So it’s a very positive and humanistic thing.
An engine for the imagination.
AI Maturity Model | Shimin Ooi
Building on the AI theme, my partner Shimin Ooi shared this chart, which explains why—seemingly all of a sudden—everyone is talking about AI:
Shimin (Twitter: @shiminooi) writes:
We are seeing a paradigm shift in horizontal software built on top of AI. Why is this happening now when AI is not a new concept? The past few years have generated better AI and the speed of improvements in AI has continued to accelerate. There are also now startups providing valuable infrastructure to the rest of the industry (e.g., OpenAI, Stability.ai, Hugging Face, etc.) instead of incumbents owning the infrastructure as well. These two factors have enabled AI to quickly create value and create disruption in verticals where there are clear software use cases and no strong incumbents.
We will likely see sustained innovation in application software tools that will eventually become more and more embedded in businesses’ DNA. For example, one will no longer need to hire a marketing agency to create a customized ad for your product—AI can help create that in less than a minute. The business applications that will likely see the most and the quickest adoption in AI will be applications with highly paid, manual, and repetitive tasks such as content marketing, sales emails or ads / image generation. We might not see humans be completely removed from the loop but businesses would gain significant efficiency from having AI complete at least 70-80% of the task.
Just as AI can amplify creativity, AI can amplify productivity. We see this in many of the tools that give writers and marketers superpowers, like Jasper.ai, Copy.ai, and Lex.
Model Size | Erin Price-Wright
My partner Erin Price-Wright shared this chart on the size of transformer models:
While GPT-3 came out less than three years ago with ~200 billion parameters, the new GPT-4 has ~1,000,000,000,000 (a trillion) parameters. Erin (Twitter: @espricewright) explains:
In 2017, Google published the seminal paper “Attention Is All You Need,” co-authored by Aidan Gomez, founder of Index portfolio company Cohere.ai. This ushered in the era of transformer models, which has radically accelerated the pace of development and adoption of Artificial Intelligence for increasingly complex tasks. The size of these models has been increasing exponentially since then. With each new model released comes a step change in it’s “humanness”—e.g. the model’s ability to understand context, generate novel ideas, recall information, or construct a complex argument.
But larger models also come with a cost—literally! Training a model with hundreds of billions of parameters can cost several million dollars. And if you want a model to stay up-to-date with current events and maintain its quality over time, you need to re-train the model on a somewhat regular basis. This means that only a small number of companies will be able to maintain these powerful models, sometimes referred to as Foundation Models. Similar to the rise of large cloud providers like AWS, GCP, and Azure, we believe that this trend will lead to a small number of AI providers, like OpenAI and Cohere, and most products and companies will use these models out of the box rather than try to build their own.
This is an exciting proposition. The cloud providers that Erin names made it dramatically easier to build with technology; they provided essential infrastructure for others to build. The same will be true of the AI providers, with companies building on top of their (expensively-trained) models.
My partner Katharina Wilhelm, who is based in Berlin, shared this chart of average monthly wholesale electricity prices across EU countries:
Yikes. Europe has been going through an electricity crisis for the past few months. Kathi (Twitter: @kathi_wilhelm) writes:
In the last few weeks, I’ve seen people wrapped in blankets in their offices, news that Berlin swimming pools will operate on lower temperatures, lighting on all sights/government buildings shut off, and people chatting about what generator to buy to be prepared for a potential blackout.
Who would have thought that in 2022 people in developed countries have anxiety and concerns over energy consumption and availability? While this is horrible, it has also produced an immediate shift towards awareness around energy-saving technologies and conscious consumption. Together with the broader sustainability challenges we face, I hope that this will create a more mindful consumer and innovation out of necessity.
The soaring electricity prices in the EU are largely the result of natural gas prices skyrocketing, a consequence of the war in Ukraine. Natural gas now costs about 10 times what it did a year ago. Electricity prices are directly tied to the price of gas, and have thus suffered.
As Kathi points out, a silver lining of this crisis could be growing awareness around the need for new solutions. On a related and more optimistic note, here’s a chart that Edward, a Digital Native reader from the Netherlands, sent me last week:
The blue lines show expert estimates for solar capacity growth, while the red line shows the actual growth. We consistently underestimate the speed of solar.
Final Thoughts: Thanksgiving
Since tomorrow is Thanksgiving, I thought I’d end with a timely chart. Here’s a chart of Black Friday e-commerce sales in America:
In 2022, e-commerce is expected to become a $1 trillion market for the first time.
One final note on Thanksgiving with one non-tech-related chart. This holiday is about spending time with loved ones, and it’s important to recognize and appreciate that valuable time together. The chart below has been making the rounds online, and in many ways it’s a quite sad chart.
The chart tracks who Americans spend time with by age. Time spent alone grows throughout your life, peaking in old age. Time with family and time with children each peak early and then decline. For 30 years of our lives, we spend 3x as much time with our coworkers as we do with our families and our children combined (though I’d be curious how remote work changes this).
A related article that’s worth reading is Tim Urban’s The Tail End, in which he calculates that by the time he graduated high school, he’d spent 93% of the time he’d ever spend with his parents.
These charts and statistics can be sobering, but they can also be an important reminder to enjoy the time spent with family and loved ones this holiday.
Happy Thanksgiving, and see you next week! 🦃
Related Digital Native Pieces
Last week’s 10 Charts (my 10 picks)
Thanks for reading! Subscribe here to receive Digital Native in your inbox each week: