Space Elephants Across the Universe: Why It’s So Hard to Understand AI
Image from Midjourney of course.
The Original Parable
Maybe you’ve heard the parable of the blind men and the elephant. It’s about a bunch of blind guys who are invited to touch an elephant and describe it. Because the elephant is so large and various parts feel so different, the blind folks cannot agree on what they are touching. In some versions, they come to blows over their different impressions. You may have seen pictures like this illustrating the lesson:
It’s not a bad metaphor to illustrate why it’s so hard to discuss AI in early 2026.
There are people out there today who are using the Gemini 2 free plan on a 6-year old Android device on spotty internet with no prompting experience. These folks have heard that AI can solve any problem, replace any human, and do it instantly. They give it a shot, are underwhelmed, and conclude that AI is overhyped.
On the opposite end of the spectrum, there are people using the latest desktop computer with 5 external monitors on a 10GB fiber line who instantly adopted Claude 4.6 Opus when it was released, have built multi-agentic frameworks to write code in parallel, and whose area of expertise is well-served by current AI coding tools. To them, current-gen AI models are not just competent and under-hyped, they represent an existential risk to their livelihood.
You probably fall somewhere between these extreme stereotypes. All of these variables (hard, connectivity, model choice, prompting experience, and problem domain) help to explain why sober-minded people might use the same technology and draw opposite conclusions about its utility.
That would make it difficult enough, but the situation we find ourselves in today is much more confusing.
The Original Parable No Longer Works
In the original parable, there is a single elephant that is being touched by several blind men. Despite their different points of reference, they are still observing the characteristics of a single mammal at a single point in time. But AI technology today is evolving so rapidly that this doesn’t really work.
Instead of a single elephant being investigated by a team of patient blind men, we have to conjure a far more bizarre situation to understand what’s happening today. The easiest way I can think of making the point is to talk about space elephants and the speed of light.
Buckle up, it’s about to get weird.
Space Elephants Across the Universe
A real photo? C’mon. It’s 2026. That’s all AI slop, friend.
Imagine that instead of a single elephant and a single pack of handsy blind men, we have a dozen elephants that are placed at random across planets outside of our solar system, but within 50 light years of earth. Each elephant is accompanied by a single blind man. Let’s assume they can survive indefinitely on their planets.
Just like in the original parable, each blind man has been instructed to describe what he is touching and send their findings to you. Let’s assume you have the above pictured radio telescope array to receive their messages.
You haven’t been told what the blind men are touching. Your goal is to piece together the fragments of their descriptions and puzzle it out. But because these planets are very far from you, at different starting distances, in constant motion, and each planet has only a single blind man investigator, the task is much harder. Let’s see how this might play out.
The nearest planet to earth outside of our solar system is Proxima Centauri, which is 4.2 light years away. Let’s say one of the elephant / blind man pairs gets lucky and is placed there. You would receive their transmission first because of how close they are.
But as soon as you had received the report, it is hopelessly out of date. The elephant and the blind man are now 4.2 years older than they were when they sent the message. And if you respond, 8.4 years will have elapsed since they sent their first report. For some teams on extremely distant planets, by the time you get their message, both the sender and the elephant will have died. This is the situation that we find ourselves in today with AI.
We are all trying to describe a highly complex phenomenon with limited information and arbitrary constraints placed on our ability to comprehend it.
Cutting Through the Noise
If you’ve gotten this far, you’ve probably bought the argument that it’s very difficult for any one person to comprehend the details of what’s happening with AI. There’s just too much happening too fast.
But even if you can’t know all the details, it seems fairly reasonable to me that AI is pretty scary and it’s getting scarier at an ever-increasing rate. AGI or ASI could exist right now – the very moment you are reading these words – but it might just be happening elsewhere to other people and hasn’t reached you yet. You wouldn’t even know that you don’t know until that information crashes into your reality.
In the meantime, posts on Hacker News and Reddit from other people will seem increasingly out of touch with the reality you are experiencing.
In my next post, I’ll cover what I think a rational actor can actually do about any of this, but for now, I’ll just leave you with the image of relativistic space elephants to convey the point that making predictions about AI is really damn hard right now.
How To Become Wildly Successful in Business
I recently finished reading The Thinking Machine about Jensen Huang, Nvidia, and the development of AI. It reminded me of the biographies of two other hyper-wealthy capitalists: Chernow’s Titan: The Life of John D. Rockefeller, Sr., and Nasaw’s Andrew Carnegie.
I was mulling over the similarities between these industry titans and realized that their stories illustrate exactly what steps you can take to become wildly successful in business. For those familiar with D&D character building: the key is to become a glass cannon. For those not quite so nerdy, it’s Mark Twain’s famous quote: “put all your eggs in one basket, and then guard that basket.”
The catch however, is that while it’s straightforward to actually take these steps, they don’t guarantee success. Like a surfer who paddles to a part of the ocean where good waves normally form, there’s no guarantee that you’ll catch an amazing wave. Conversely, of course, the surfer who stays on the shore has no chance at all.
So here’s the formula in a nutshell. I’ll break it down in greater detail below using examples from the lives of Jensen Huang, John Rockefeller Sr, and Andrew Carnegie:
Find a business activity that you intrinsically enjoy doing.
Do that thing to the exclusion of almost everything else in your life.
When opportunities present themselves to diversify your time, energy, or money, ignore them and double down on your thing.
Get lucky.
Do Something You Enjoy
It’s pretty hard to imagine anyone pulling 90-100 hour weeks month after month, year after year if you don’t intrinsically enjoy doing the work. Some people have insane self-discipline, but in the contest between people of approximately equal skill, one of whom just can’t get enough of doing something, and one that has to force themselves to do it, it’s pretty obvious who will win out over the long-term.
Jensen Huang gravitated towards electrical engineering from a very early age. People that knew him during his brief time at AMD and then LSI recounted stories of his single-minded focus on becoming one of the best chip designers not just at his company, but the industry. His work ethic was a force of nature.
Similar accounts are told of both John D. Rockefeller and Andrew Carnegie. And it’s worth noting that during the decades they grew up, a “standard” work week was 60-70 hours. Contemporaries describe them as out-working the most diligent employees, staying later, waking earlier, and just generally out-grinding them.
John D. Rockefeller learned the ropes at the trading house of Hewitt & Tuttle where he developed a passion for accounting and negotiation. He was later able to leverage those talents when he started Clark, Gardner & Company where he first started trading and refining oil.
Carnegie started his career as a telegraph operator before moving on to the Pennsylvania Railroad company. When he wasn’t working, he was reading and learning. First about telegraphs, then railroads, and later about steel manufacturing. Carnegie’s great business passion was learning and improving.
So, paradoxically, to become wildly successful in business, you first need to identify a business activity you love doing. And then you need to do it for almost every waking second of your life for many years.
You could insert a glib quote here about hitting 10,000 hours of experience, but I don’t think there’s anything that predictable. Anyways, if you’re doing this right, you won’t care whether it takes 10,000, 15,000, or 30,000 hours to become an expert, you’ll be having fun the entire time.
For Jensen, it was designing circuits, for Rockefeller, it was accounting and negotiation, and for Carnegie it was the thrill of learning something new. Each spent years in their early careers doing almost nothing else but learning their craft.
Do That Thing to The Exclusion of Everything Else
To become wildly successful at anything of importance, you need to put in the hours. But human lifetimes are pretty short, so if you want to achieve domain expertise while you are still young and healthy enough to leverage it to achieve great things, you can’t afford the luxury of diversification.
Jensen Huang worked subscribed to the 996 lifestyle (9AM to 9PM 6 days per week) while at LSI. He was responsible for designing novel chip architectures and delivering them to vendors. By the time he joined the founders of Nvidia, his life had consisted mainly of designing computer chips and running teams of technologists.
Rockefeller learned his craft at a trading house in early career, but then struck out on his own. He quickly became the expert among the 3 founding partners and maneuvered them out of the business so that he could move things more quickly. His days consisted of nothing but learning markets and negotiating with vendors and transport companies.
Andrew Carnegie taught himself to interpret morse code just by spending time among telegraph operators. At the Pennsylvania Railroad company, he taught himself the then-legal skill of insider trading and when he founded the Keystone Bridge Works company, he taught himself how to optimize the Bessemer Process to improve the efficiency of producing low-carbon steel. He only married at the age of 51. When he wasn’t working, he was reading books and learning about new technologies and management practices.
While all three men did marry and have children (eventually), their focus on work was a meaningful tradeoff against other life paths. It’s not that you can’t have a healthy marriage or hobbies if you want to become wildly successful in business. For example, Rockefeller was an avid golfer later in life and Jensen apparently has a quirky interest in teppanyaki. But if you want to get good at a skill quickly, it will crowd out lots of other things you could have spent your time on.
Never Diversify, Always Double Down
If John D. Rockefeller Sr. had followed standard portfolio development advice and diversified his stock holdings outside of the Standard Oil Company, he never would have become so wealthy. Instead, it was well understood at the company during his tenure that if you wanted some liquidity, you could always sell your shares to John and he would buy them for the market rate. He did this for decades and without regard to the prevailing sentiment about the future of the company. His belief in himself and the value of the products the Standard Oil Company made was unwavering.
Carnegie leveraged his connections with magnates in the railroad industry to make a small amount of money early in his career that he then deployed to finance the Keystone Bridge Works which would eventually become US Steel. Instead of spending the profits from his earlier ventures, he ploughed those proceeds into acquisitions of competitors, starting with the Homestead Steel Works. The name “US Steel” was later chosen because he had acquired so many competitors he needed a generic name to encompass the resulting corporate entity.
Like Carnegie and Rockefeller, almost all of Huang’s wealth has been generated by a single company, in this case Nvidia. While we know less about Huang’s personal finances than we do about the late Rockefeller and Carnegie, he has been the chief executive officer of Nvidia for 32 years, making him one of the longest-serving CEOs in the technology industry by a wide margin. He has spent the literal majority of his life doing nothing but building and scaling Nvidia. The whole premise behind his push for CUDA and its associated software libraries was a massive doubling down on parallel computing that was viewed by many industry experts as foolhardy and misguided. But Jensen had a vision and he doubled down on it, even when that vision seemed likely to wreck his company and career.
For me, the crazy thing about these stories is the scale of these men’s focus.
Side quests get more and more appealing the more experienced and renowned you become. You might be lucky to scrape together an unpaid internship in your chosen activity when you’re in your early 20s. But when you are 45 and have developed a reputation as the best hardware entrepreneur in the industry, you probably get all-expenses-paid speaking gigs to the island of Fiji with $250,000 honorariums. And while this may seem like an obvious distraction from your laser-focused pursuit of your activity, others will be more subtle.
Standard Oil could have become a broader commodities broker, US Steel could have expanded to producing other metal alloys, and Nvidia could have branched out into services or chip fabrication. Instead, each business leader chose their niche and at every turn.
So to become wildly successful in business, you have to avoid the siren’s call of diversification at every turn. Even if that means you have to literally create new markets for your products.
Prior to Rockefeller’s refinery optimizations, most refiners treated gasoline as a waste product and dumped it into rivers and ponds just to get rid of it. One of his key insights was that refinery by-products could be sold for other uses.
By bringing down the cost of hardened, low-carbon steel, Carnegie wasn’t just able to build stronger and longer bridges for Keystone Bridge Works, he created entire military defense applications like metal battleships.
Huang’s expensive gamble on CUDA hardware and associated software libraries enabled advances in neural nets to produce the modern AI industry.
These entrepreneurs didn’t just double down: they literally created the reality their businesses needed to be successful.
Get Lucky
Finally, and most critically, you have to get really lucky. Rockefeller, Carnegie, and Huang all identified a business skill that they intrinsically enjoyed. They put in the time to become world class in those areas. They resisted the temptation to diversify, and instead spent decades doubling down on their previous wins.
But these investments are fundamentally fragile. Without a substantial amount of luck, the stories of these men’s lives would likely have been forgotten after they doubled down one too many times and lost everything in a bad turn of fate. That’s what happens to almost everyone throughout history. We only know Rockefeller, Carnegie, and Huang’s names because they did what is nearly impossible: they made huge sacrifices, put all their chips on black, won, put all their chips on red, won, and then did that another 20 times.
Rockefeller entered the oil market against the preferences of his then-business partners who didn’t understand it and felt it was too volatile and immature. But Rockefeller got extraordinarily lucky with his timing. After the Civil War, oil prices cratered as the northern industrial economy rapidly demobilized. The Pennsylvania oilfields around Titusville went from boom to bust and margins for refiners collapsed. Rockefeller’s first lucky break was in his competence negotiating (often brutally) for the absolute lowest rates for transporting crude and refined oil products. His years of honing his accounting and contract negotiation skills combined with a bust among pre-war drilling companies gave him unprecedented leverage to out-extract and out-transport his competitors.
Although the Bessemer Process predated Carnegie’s entrance into the world of steel manufacturing by more than 20 years, that interim time had allowed the technology to become more widely deployed and ripe for innovation. Carnegie streamlined and vertically integrated the manufacturing of steel at an unprecedented scale during a period of rapid economic growth following the American Civil War. His deep contacts in the railroad industry helped his company secure contracts even during the economic depression from 1873 - 1879 and his connections in Washington helped US Steel receive favorable tariff protection. If even one of these ingredients hadn’t been present, it’s feasible that US Steel would never have been created.
Like most startup companies, Nvidia nearly died several times in the early years due to bad product releases, economic conditions, and poorly-timed strategies. But Jensen got extremely lucky that his expensive investment in CUDA hardware and software was timed perfectly for use by cryptocurrency miners and neural net advances. Without those lucky breaks, the company may well have been relegated to the dustbin of history along with other contemporary graphics card companies like 3DFX.
To continue the surfing metaphor, you have to get exactly the timing and weather if you want to catch the perfect wave.
You Feelin’ Lucky Punk?
Most people who truly fathom the costs associated with this path eventually shy away and choose to compromise. That’s not a failure of personality, but a healthy understanding of their values.
But for those that want to become wildly successful in business, the steps are pretty easy to explain. The question that you have to ask yourself is whether you’re feelin’ lucky. If you are, go ahead: find a skill, focus on it, double down, and roll the dice.
A Frustrating Adventure Trying To Design A Logo With AI
I started my career in tech as a product designer. Since then, AI has fundamentally changed the way most design work gets done. I’ve been having fun following along in my limited free time. For instance, last year, I taught myself Figma.
Most tools today give designers the ability to ideate alongside AI models. In general, I haven’t found them to be very useful, but it’s still early and I always figured that I might just be a n00b. After all, I haven’t been a full-time designer in 10+ years, so every time I tried using one of these AI design tools, in the back of my mind I couldn’t help but think “your inability to get what you want might be user error.”
One area of design in particular that I’ve been disappointed in AI output has been in the logo design department. I’m not sure exactly why AI logo results have been so lackluster, but I had this nagging sense that maybe if I spent an hour or two really focusing on this, I might be able to get much better results.
So I took some time on MLK day to go deep on this and see whether present AI capabilities really are lackluster or if I’m just doing it wrong.
After testing 13 AI design tools across several different prompt types, I’m forced to conclude that AI probably just can’t do a good job of creating logos at this point.
I have some ideas as to why this might be the case and I discuss those at the end, but before we get there, join me on a tour of some wonderful AI slop!
The Test Product
For this experiment, I decided to help a friend out with an app that he’s building. It’s a niche app in the heavy manufacturing industry. He’s calling the product PAX which is an acronym for Power Asset Exchange. Basically, his company connects power generators (utilities and middlemen) with used gas turbine parts.
I spent a couple of minutes iterating a prompt with ChatGPT and used this throughout all of the testing.
Prompt
Design a simple logo for a company called Power Asset Exchange (PAX) that helps power generators by providing them with used gas turbine parts. The company operates in the heavy manufacturing industry.
The logo must be simple. This is the most important constraint.
The logo must have a square 1:1 ratio.
The logo must be visually distinctive in both black and white and color.
The logo must be easily identifiable at the size of a favicon (16x16px).
The logo must be conceptually relevant to the brand.
To reiterate, the logo must be very simple - fewer lines, fewer themes, fewer shapes. The ideal logo has only a couple of edges and shapes and even fewer colors.
Attempt 1: Free Online Tools
I did a quick pass through some of the free logo design tools and the results were pretty bad, so I didn’t spend much time trying to iterate. My current hypothesis is that these tools are able to offer a free version of the product by using older, cheaper models and they aren’t using evals to actually measure the quality of outputs.
Looka
Logo.com
Adobe Free Logo Maker
Design.com
Logomakr
Looka
Looka doesn’t do a terrible job, but all of the logos feel like they should be for a small town CPA firm, not an innovative tech manufacturing company:
Logo.com
They paid a lot for that domain, but don’t appear to have invested much in differentiating the product. Same complaint as Looka:
Adobe Free Logo Maker
Nothing much to see here. I got several of these, but at this point, I strongly suspect that the system prompt and model that each of these companies are using is very similar:
Design.com
These are actually pretty good compared to everything else I’d seen so far. You could imagine putting in some time to tweak and simplify one of these into a working concept. Still not excellent, but getting warmer:
Logomakr
The guided prompt builder approach here looks promising, but the output was disappointing. It’s not terrible, it’s just not very inspired. The tool promises that you can go further and iterate, but the starting point doesn’t inspire confidence:
Attempt 2: Dedicated Design Tools
Maybe tools that are actually built for designers will perform better than the free, one-shot, off-the-shelf variety. Here I considered 5 products, some of which you’ll probably recognize:
Canva
Lovable
Recraft
Midjourney
Figma
Canva
Canva does an okay job. It created a couple of design ideas, I picked the one that looked the best:
I have no idea what the graphic is supposed to be and it has some inexplicably-complicated stuff in the middle. The font is okay, the border looks dated. Overall, I wasn’t impressed.
Recraft AI
I heard about Recraft from a friend and was pleasantly surprised by the results. Still not super inspiring, but these are at least simple and would be easy to amend:
Lovable
I got a bunch of ideas out of Lovable, but this seemed to be more of a quantity over quality approach. They were at least simple, though:
Midjourney
I know, I know, Midjourney is normally used for pictures, but why not give it a shot as well?
The black and white aesthetic is nice, but I have no idea what these designs are trying to convey. There’s no indication that any of these are logos for an energy company. Oh well, it was worth a try.
Figma
Figma is the 800 lb gorilla in the design space these days. Like all modern design tools, it has a generative AI mode. I should note here that Figma’s generative AI feature appears to work by creating SVGs rather than rendering pixels, but since it IS a purpose-built design suite, I felt like including it was fair.
Somewhat hilariously, Lovable and Figma gave me nearly identical designs:
I’ll hand it to Figma that this design is indeed simple, but there just isn’t much to connect it to a power generation company, the full product name, or the acronym.
Attempt 3: Just Use The Leading-Edge Models Directly
So maybe the solution is to go directly to the best leading-edge models and do some prompt engineering to get better results. I have paid subscriptions to all the major models, so I might as well get my money’s worth, right? Here’s what I tried:
ChatGPT 5.2
Gemini Nanobana Pro
Claude 4.5 Opus (note that Claude doesn’t have native image generation capabilities, but I figured I’d give it’s SVG generation a try
Claude’s attempt was a complete failure, but that’s reasonable given the product’s constraints.
ChatGPT and Nanobana’s results aren’t terrible, but they lack consistency or a unifying concept. For instance, why is the ChatGPT logo not symmetrical? Nanobana’s attempt looks neat, but the “P” looks a lot like an “F” and I don’t know why.
Attempt 4: Starting From Examples
Up to this point, I had been using words to describe the prompt design, but what if I uploaded some reference logos in addition to the prompt?
The industrial manufacturing sector isn’t a bastion of cutting-edge mobile and web UI design, but I was able to quickly locate a few clean-looking logos:
I uploaded 2 of these to Figma just to see whether the results changed much. To my surprise, I got a borderline-usable output:
I thought, “great! It’ll be even better when I add in the other data points.” So I uploaded the remaining two and … we’re back to unusable:
Conclusion
It was pretty fun trying a bunch of different prompting approaches, tools, and comparing notes. During a normal work week, I don’t have the time to really go deep to differentiate whether some AI output is garbage and I need to do the work myself or I’m just bad at prompting.
My takeaway is that at least in the domain of logo creation, AI isn’t all that useful right now.
“But George,” you might say, “there are dozens of ideas that you got for nearly free! Surely this is an improvement over what came before!”
And I would agree with that sentiment, I just don’t think it’s a significant improvement. Here’s why: actually using one of these AI-generated ideas for a real business would take a significant amount of rework.
For instance, scaling a logo down to 16x16 for use as a favicon requires more than just resizing a high-res version of the logo. To get something clean and discernable at the smaller size requires a rethink of the design elements into individual pixels.
Similarly, most businesses need a bunch of different aspect ratios to be rendered for different social media platforms, trade shows, and marketing materials. I did this a lot for my two companies and it’s a surprisingly creative process. For one ad format, you may need to isolate an element and render it alone. On a newsletter masthead, you may need to create a completely horizontal version of the logo that was designed to be square. Businesses need clean, organized underlying assets to rapidly create design resources like these.
So, while I think AI models can probably create decent fodder for brainstorming, I think we’ll need to wait for better underlying models before we can seriously talk about automating designers out of the logo creation process.
Why is B2C User Acquisition Broken?
Backstory
Right after graduating from college, I started my first company, Skritter, with my two best friends. It’s an app that helps students learning Chinese and Japanese write characters. We didn’t do anything special to get users: we created content on our blog, hosted a forum, learned about a new thing called SEO, and kept releasing new features. The company grew little by little via word of mouth. The company is now old enough to vote and is still growing.
In 2013, we all stopped working on Skritter to start our second company, CodeCombat. It’s an app that helps middle schoolers to code. We were part of the W14 YC batch. At CodeCombat, we became experts at getting onto the HackerNews front page, but most of the lasting user growth was viral. Learning to code was hot in 2014 and our product was eventually pretty good. Today, the company is several times larger than Skritter and continuing to grow.
After CodeCombat, I was experiencing burn out and decided to work a corporate gig for a while. But I learned the hard way that it’s really hard to scratch my entrepreneurial itch at a corporate gig. So I did what any sane father of 3 does: I started building some side projects in my free time.
It can be fun to build for yourself, and I’ve done it plenty of times, but it’s a lot more fun to build something that other people want. So, rather than just building the 10,000th bespoke Quantified Self app for myself, I set out to find the overlap between 1) things I want to build and 2) things that other people might want to use.
This time around, I wasn’t interested in creating another startup. I had pretty modest goals: build something that other people find valuable and maybe generate a little income on the side.
Little did I know that at some point during the intervening decade, B2C user acquisition broke.
My 3 B2C Attempts
I had a couple of ideas that I was passionate about and I thought might resonate with other people:
Never Apply
My idea here was to help people find jobs through their network. $1,000 in LinkedIn ads later and there was no engagement at all. I drew what seemed then like the natural conclusion: no one wanted this because my idea or execution were bad. No big deal, running companies has taught me that most of the stuff you try doesn’t work. Onwards to the next thing!
Here’s what the homepage looked like:
Talk with Sage
I’ve struggled with depression my entire life and I thought there might be a niche in making a much better agentic AI chatbot therapist. I did some organic Reddit content creation and also bought ads. Here too, I got literally 0 engagement. No signups, no support requests, no rants about pricing. Absolutely crickets.
Like Never Apply, I thought, well, I probably just didn’t clearly express the value-add or find the user group who would actually use this thing. Another case of poor execution or market segmentation. But it was mildly interesting to me that I’d had two absolute flops. You sort of expect things to fail when building new stuff, but getting zero signups or signs of life was a bit extreme.
Here’s what the homepage for this one looked like:
Family Caller
My grandmother was diagnosed with early stage dementia last year and her care has been challenging for my family. I offered to build this one to help my mom. I felt a lot more confident about this one because I personally know a couple of family members caring for elderly folks. They all agree it is hard, draining work. My initial run of Meta ads showed some of life, so I decided to invest a bit further.
I doubled down, iterated the product messaging and dove deeper into the actual product build. But despite that initial ads-based interest, there was almost zero engagement with the product.
And this is when I started to think that something systematic might be broken with the ecosystem, not the products I was testing. Family Caller is still live, here’s a screenshot of the homepage:
None of these ideas seemed obviously stupid. I helped to design each of the pages with a designer who is a lot better than me, and Scott helped me out and built the backend for Family Caller. True, each of them had noticeable rough edges, but I showed the pages to some friends and family without first telling them I’d made them, and the feedback was generally “this seems legit.”
Broken B2C User Acquisition
At this point I need to stress the problem: there was almost no engagement at all. It wasn’t that what we built didn’t work from a business perspective. We instrumented the apps pretty well and what we saw is that people (or maybe just bots) were looking at the pages, scrolling, and tapping, but almost nobody actually used the products.
Here’s all the stuff we tried to figure this out:
Conducted in-person UX testing. At first, I thought maybe the pages were misleading and that people were signing up thinking that each service did something else. That might explain why accounts were getting created and then going silent. So I did some UX testing with live humans: they had no problems navigating the sites or explaining what each product did.
Installed Posthog for screen recordings. Dead internet theory 101 says that most activity on websites are crawlers and bots. So I installed Posthog to give me screen recordings of sessions. You can’t tell definitively from a recording whether it’s a person or a bot, but if these were bots, they were super inefficient and idiosyncratic. I concluded that either they were sophisticated bots built to closely emulate somewhat weird human behavior or actual humans.
Installed Cloudflare’s Turnstile to screen for bots. This only partially worked and introduced noticeable bugs. Via Posthog user sessions, we were able to see that a fair number of sessions were getting Turnstile errors that broke the site for them. These weren’t “you’re a bot, no more pages for you” notices, these were “Turnstile broke everything” errors. We ended up uninstalling it.
Emailed users. I had legit-looking email addresses and I figured that if someone was willing to part with their email, they must have some shred of interest in the product. So I tried a couple of email tactics. I reached out asking for feedback. I offered first unpaid and then paid usability testing sessions. I emailed to “check in” to see if they needed help completing onboarding. I sent all of these emails personally from a Google workspace account 1 by 1 to maximize delivery. Out of the couple hundred I sent, I got one generic reply.
Installed a chat support bubble. We installed a “chat with us” feature for a while and we got … 0 chats, despite manning the chat threads myself and trying hard to ensure the users knew I wasn’t an AI agent. I guess I failed the reverse Turing Test.
After all this, I was even more convinced that B2C user acquisition might just be broken compared to a decade ago when my cofounders and I were running CodeCombat. But why?
Nobody Buys Software Anymore
I track my finances religiously and am responsible for managing our spending categories. A couple of years ago, I mentioned to my wife that even though she and I are always worried about paying for software subscriptions that we don’t use, we buy almost no software these days. Here’s a full dump of all the software that we’ve bought in the last 12 months:
Steaming entertainment: Netflix, Hulu, and Disney+
Big Tech services: Microsoft Office, Amazon Prime, Youtube Music
One-offs: TinyBeans (yearly kid photo storage subscription), Partiful (nice party invites, only 1 month for a party we hosted), Quicken (yearly subscription)
So apart from streaming entertainment (which I would argue is more akin to cable television than software) and services from FAANG / MAANG, we just aren’t buying much software.
Next, I asked a couple of my friends about this. The way I phrased the question was “apart from FAANG and streaming services, do you buy software?” The answer seems to be “not really.”
So what happened? I remember a decade ago, I bought all sorts of niche software. I had a Zeo headband with a monthly data analysis subscription, I tried the paid tier on Strava, I bought the pro plan on AllTrails, etc, etc. I wouldn’t say that I was a software spendthrift, but I certainly paid for more B2C software than I do today and I arguably had fewer problems to solve in my life. (nothing beats kids for creating life problems!)
Then I had an idea: what if this isn’t just me and my little social bubble? What if people in general are just not buying much software anymore? That would be kinda weird, but I had an easy way to sanity check my idea.
It’s Not Just Me, YC Is Funding Mostly B2B Startups Now
YCombinator has a publicly-available list of all the companies that they’ve funded: YC Startup Directory. They are funding a lot more B2B companies now than B2C:
YC has funded 5,407 companies. 2,685 are B2B, only 838 are B2C. That’s a telling statistic on its own. The single biggest startup accelerator in the world has funded approximately 3x as many B2B companies as B2C.
But if you take a look at the time trend, the data gets even more stark. I did some tedious scraping and number-crunching from public sources for you. The percentage of B2C companies that YC funds per batch has fallen from around 30% in the early 2010s to around 7% in recent cohorts.
And that number actually understates the drop. If you actually dig into the B2C companies in the most recent cohorts, most don’t monetize their products directly. IE, people aren’t paying for them, they’re ad-supported, VC-backed, or funding themselves via affiliate revenue.
So, putting this all together:
My own little side experiments have proven that user acquisition for B2C products is not working for hobbyists.
My peers and I no longer buy much software directly.
The YCombinator Startup directly provides strong evidence that B2B is where venture-scale investments pay off, not B2C.
Despite these trends, the tech industry is bigger than ever as measured by public markets and VC funding:
So what’s going on?
Some Hypotheses
After thinking about this a bit, I have a couple of hypothesis that might explain this trend, but would love to hear from readers to see whether I’m overlooking some more obvious explanations:
Software ate the world, now there’s nothing left. Andreessen may have been right that software is eating the world, but perhaps now there’s no world left. If there are 50 note-taking apps, the 51st is going to struggle to find users, even if it’s demonstrably a lot better. Maybe the problem is literally that there’s too much software.
AI bots have made user discovery impossible. Large language models are flooding product communities faster than they can be detected and removed. It’s possible that this asymmetric advantage has essentially spoiled the ability of new products to find their users. Maybe with time, defensive tactics will catch up and things will be like email spam in the early 2001s: a major problem until it isn’t any more. But for now, AI could just be ruining things. This trend applies to B2B as well: cold emailing, content marketing, and direct phone sales used to work and are no longer viable strategies. The one saving grace of B2B is that business owners are real humans that have real needs and you can sometimes talk to them.
Growing distrust of tech in general. Maybe people are turning away from tech-driven recommendation engines like ads and feeds as part of the tech backlash that kicked off in the mid-2010s. You could imagine a world where average people have just stopped trusting that they’ll get good value from any digital purchases after they struggle to cancel the subscription on their fitness app or the accumulation of dark patterns in their social networking app of choice leaves them with a bad taste in their mouth.
The middle class is finally gone. We’ve been hearing the story about wage inequality for a long time, but maybe things have finally become bad enough that mass-market apps that aren’t explicitly tailored for the top 1-10% just aren’t viable. Anecdotally, I know quite a few underemployed and unemployed people my age that are finding it extremely difficult to get jobs or save money, so this one too seems at least plausible.
Of course, it could be all of these things plus 2-3 more that I’m missing. What do you think: why is B2C broken for new startups?
Books That I Read in 2024
Nonfiction
★★★★★ Different: Gender Through the Eyes of a Primatologist - changed the way that I view the behavior of myself and my young children.
★★★★★ Not the End of the World: How We Can Be the First Generation to Build a Sustainable Planet - some much needed optimism about climate change and our ability to change the world.
★★★★★ Humankind a Hopeful History -reminded me of the book “Factfulness,” it’s a good read for anyone feeling gloomy about the state of the world.
★★★★★ Outlive: The Science and Art of Longevity - lots of people have already learned this stuff, but I haven’t been paying attention to nutrition or longevity research and it was a fascinating read.
★★★★★ The Dawn of Everything: A New History of Humanity - a good rebuttal to “Sapiens” and “Better Angels of Our Nature.”
★★★★☆ The Hidden History of Guns and the Second Amendment - well-researched argument that gun rights and slavery are tightly entwined in ways that most people don’t want to talk about.
★★★★☆ Why We Fight: The Roots of War and the Paths to Peace - apparently we don’t start wars over resource scarcity which was a surprise.
★★★★☆ Judgment at Tokyo: World War II on Trial and the Making of Modern Asia - I learned about the Nuremberg Trials in school, but never learned about the Japanese equivalent. This was a long but fascinating read.
★★★★☆ The Little Book of Aliens - I don’t know where I learned that humanity had done a lot of searching for aliens, but this book disabused me of that fact.
★★★★☆ Freezing Order: A True Story of Money Laundering, Murder, and Surviving Vladimir Putin's Wrath - this and “New Cold Wars” pair well together if you’re interested in an analysis of US / Russian relations.
★★★★☆ The Survivors of the Clotilda: The Lost Stories of the Last Captives of the American Slave Trade - change is never tidy and uniform, this book does an amazing job of bringing to life a transition that was always quickly and factually summarized in school history books.
★★★★☆ Blitzed: Drugs in the Third Reich - I knew that lots of drugs were legal in the early part of the 20th century, but didn’t realize that the Third Reich was meth’d up and that it may have been responsible for some of their early military victories.
★★★★☆ Four Thousand Weeks: Time Management for Mortals - I had been avoiding working on some important life projects because they didn’t feel enjoyable. This book explained to me why that is and why it’s imperative to do the work anyways.
★★★★☆ The Emperor of All Maladies: A Biography of Cancer - if you thought that cancer was monolithic, you should read this and marvel at how our best enemy really is ourselves.
★★★★☆ The Rise of Theodore Roosevelt - so much has been written about TR’s presidency, but not as much about his adolescence and early adulthood. You can disagree with some of his policies and still respect the man.
★★★☆☆ Blight: Fungi and the Coming Pandemic - pair this with the TV show “The Last of Us.” You’re welcome.
★★★★☆ Nuclear War: A Scenario - prior to reading this book, I ranked nuclear war as perhaps the 3rd or 4th more pressing existential threat to humanity. Now I think it’s probably #1.
★★★★☆ The Price of Time: The Real Interest Story - a compelling and well-researched argument against low interest rates.
★★★★☆ The Upswing: How America Came Together a Century Ago and How We Can Do It Again - a great snapshot of historical consequence, but I fear that what worked previously won’t work now that Facebook, Instagram, and video games exist. Just the same, fascinating.
★★★★☆ The Demon of Unrest - I spent a full semester learning about the American Civil War and had never learned about this pivotal event that started it all.
★★★★☆ The Coming Wave: Technology, Power, and the Twenty-First Century's Greatest Dilemma - would have been better if it had taken more seriously some of the existential AI threats, but I still really liked it and thought the points were timely and well-articulated.
★★★★☆ Takeover: Hitler's Final Rise to Power - before reading this, I thought that the political situation in the US now was very similar to Germany in the 1930s. Turns out I was quite wrong.
★★★★☆ Smoke and Ashes: Opium's Hidden Histories - if you can get past the kinda weird intro, this is absolutely addictive read about a botanical history I’d never heard.
★★★★☆ Magic Pill: The Extraordinary Benefits and Disturbing Risks of the New Weight-Loss Drugs - the author might not be super trustworthy, but GLP-1 weight-loss medications are borderline magic and may fundamentally change the world.
★★★★☆ Of Boys and Men: Why the Modern Male Is Struggling, Why It Matters, and What to Do About It - finally a politically centrist book about some of Charles Murray’s points about gender and sex without all the conservative axe-grinding and culture war baggage.
★★★☆☆ Smoke Gets in Your Eyes: And Other Lessons from the Crematory - made me think harder about death one moment and laugh out loud the next.
★★★☆☆ The Lost Tomb: And Other Real-Life Stories of Bones, Burials, and Murder - pure entertainment, this is a real page turner. It’s what would happen if Indiana Jones were actually real.
★★★☆☆ Number Go Up: Inside Crypto's Wild Rise and Staggering Fall - I didn’t know most of the history of crypto currencies even though I’d lived through it. Good, if fairly opinionated read.
★★★☆☆ Empire of Pain: The Secret History of the Sackler Dynasty - don’t watch the TV show, read the book. This also pairs well with “Smoke and Ashes” above.
★★★☆☆ Lies Across America: What Our Historic Sites Get Wrong - high schools should be teaching from books like this not textbook propaganda. This is entertaining and illuminating.
★★★☆☆ The Man from the Future: The Visionary Life of John von Neumann - John von Neumann was truly an impressive dude. A bit scary, but also impressive.
★★★☆☆ Hunting the Falcon: Henry VIII, Anne Boleyn, and the Marriage That Shook Europe - it turns out that Anne Boleyn was a total badass.
★★★☆☆ The Wide Wide Sea: Imperial Ambition, First Contact and the Fateful Final Voyage of Captain James Cook - I’m surprised that this story hasn’t been turned into a TV miniseries already. It’s got everything you need: murder, intrigue, disease, war, sex, suicide, and ambition.
★★★☆☆ The In-Between: Unforgettable Encounters During Life's Final Moments - good reminder about mortality and the beautify of being human.
★★★☆☆ The Worlds I See: Curiosity, Exploration, and Discovery at the Dawn of AI - not as good as “The Coming Wave,” but still an interesting read for context about how current-gen AI systems were created.
★★★☆☆ Between Two Kingdoms: A Memoir of a Life Interrupted - even though we’ve made so much progress treating cancer (see “Mother of All Maladies” above), surviving cancer sucks.
★★★☆☆ Tokyo Vice: An American Reporter on the Police Beat in Japan - don’t watch the TV show, read the book. It’s a lot more realistic, complicated, and entertaining.
★★★☆☆ The Indifferent Stars Above: The Harrowing Saga of the Donner Party - there are some incredibly visceral and memorable human stories in here.
★★★☆☆ Alexander at the End of the World: The Forgotten Final Years of Alexander the Great - I’ve read quite a bit about Alexander the Great, so this wasn’t a revolutionary book for me, but still interesting.
★★★☆☆ The Trading Game: A Confession - I listened to this on Audible and the author’s narration makes it 2x as good. It’s like “Chaos Monkeys” but for finance.
★★★☆☆ Marilyn Monroe: The Biography - turns out she didn’t commit suicide and may, in fact, have been killed? Interesting story about an American icon.
★★★☆☆ Tripped: Nazi Germany, the CIA, and the Dawn of the Psychedelic Age - not as good as Blitzed, but still an interesting read.
★★★☆☆ Waco Rising: David Koresh, the FBI, and the Birth of America's Modern Militias - now that I’m a Texan, I really enjoyed learning the history of the city directly to the north.
★★★☆☆ The Molecule of More: How a Single Chemical in Your Brain Drives Love, Sex, and Creativity - I started this one for the neurobiology, but ended up taking away some interesting lessons about the brain chemistry of people in different political camps.
★★★☆☆ Dealers of Lightning: Xerox PARC and the Dawn of the Computer Age - the couple of depictions Xerox park being pillaged by Steve Jobs that I’d seen are wrong. It’s a quick read, so if you work in the tech industry you should probably read it.
★★★☆☆ New Cold Wars: China's Rise, Russia's Invasion, and America's Struggle to Defend the West - good, but very, very long. He makes a pretty convincing argument that the US government hasn’t done a great job responding to either China or Russia across administrations and we need to put aside partisanship and get serious about being more hawkish.
★★★☆☆ The Heat Will Kill You First: Life and Death on a Scorched Planet - great read if you live in a hot climate because the heat really is dangerous and we don’t commonly admit that.
★★★☆☆ The Woman in Me - I’m not a huge fan of Britney Spears, but it was sad and illuminating to hear about the struggles that she’s endured.
★★★☆☆ Danger Zone: The Coming Conflict with China - pair this with “New Cold Wars” to get a more balanced read on Chinese / American relations.
★★★☆☆ In the Heart of the Sea: The Tragedy of the Whaleship Essex - this is the story that Moby Dick is based on and it’s better than the novel.
★★★☆☆ Into the Clear Blue Sky: The Path to Restoring Our Atmosphere - didn’t learn a ton, but it was interesting to learn about how much more valuable it would be to remove atmospheric methane than CO2.
Fiction
Station Eleven by Emily St. John Mandel - very different than the TV show and (like almost every book), better. If you enjoyed the show, you should definitely read this.
Termination Shock by Neal Stephenson - pretty fun read, but drags a bit at the end.
The Three Body Problem, The Dark Forest, and Death's End by Cixin Liu - I devoured all 3 of these in 2 weeks. Very entertaining and better than the TV show.
Books I Started, but Didn't Finish
Are Prisons Obsolete? by Angela Y. Davis
The Woman They Could Not Silence by Kate Moore
Notes on a Nervous Planet by matt Haig
The Great Mortality by John Kelly
November 1942 by Peter Englund and peter Graves
Behave by Robert Sapolsky
The 99% Invisible City by Kurt Kohlstedt and Roman Mars
Hits, Flops, and Other Illusions by Ed Zwick
1177 B.C by Eric H. Cline
Evicted by Matthew Desmond
Plagues and Peoples by William H. McNeill
Solito by Javier Zamora
The Warmth of Other Suns by Isabel Wilkerson
Life After Power by Jared Cohen
Cloistered by Catherine Coldstream
Days of Rage by Bryan Burrough
The Small and Mighty by Sharon McMahon
The 10 Books I'm Most Excited to Read in 2025
Gray Matters: A Biography of Brain Surgery
House to House: An Epic Memoir of War
A Fatal Inheritance: How a Family Misfortune Revealed a Deadly Medical Mystery
As You Wish: Inconceivable Tales from the Making of The Princess Bride
Prequel: An American Fight Against Fascism
The Achilles Trap: Saddam Hussein, the C.I.A., and the Origins of America's Invasion of Iraq
The End of Everything: (Astrophysically Speaking)
Jesus and John Wayne: How White Evangelicals Corrupted a Faith and Fractured a Nation
Soul Full of Coal Dust: A Fight for Breath and Justice in Appalachia