š šš©š©šØš«šš®š§š¢šš¢šš¬ ššØ š«šš¢š§šÆšš§š šš”š ššØš«š„š š¢š§ šššš
John Rush
Posted on January 3, 2024
Once in a generation, we have the chance to reinvent pretty much everything.
The 5 opportunities I'm most excited about:
š. ššØš§šš«šØš„ šÆš¬ ššššØš¦š¦šš§ššš.
In the old days, we'd just watch what was shown on TV, and listen to what's played on the radio.
We had little control, mostly high-level: the channels.
Then everything went into full control. Download songs into your playlist, subscribe to the author you wanna get updates from, select movies from libraries, etc.
ššššš« š”š®š¦šš§š¢šš² šš¢š¬ššØšÆšš«šš š šššš š§šØššØšš² šš§šš¢šš¢š©šššš.
It turned out that we didn't really wanna make an effort to make decisions. We wanna have control, but also we're happy if the decisions are made for us, as long as we're happy with the results.
Today, most of us consume music, TV, YouTube, and social media this way. Most didn't notice, but the share of the content from accounts you don't follow on social media keeps growing.
ššØš®šš®šš's home page is generated mostly out of videos from authors you aren't subscribed to.
šš©šØšš¢šš²'š¬ "explore", "daily mix" and "radio" features turned on to be super viral.
ššššš„š¢š± predicts with very high accuracy the movie with its "% match".
šš§ šØš©š©šØš«šš®š§š¢šš²: take any successful consumer product and remove the need for manual decisions for the user, read the user's mind, and show what the user will end up liking as it was carefully selected by themself.
š. ššš ššØšØš„š¬ šš => šš šš šš§šš¬
One of my B2B projects is a SaaS platform for cinemas.
I was building it for a decade.
When I started, the idea I pitched to cinemas was this: you waste so much labor on simple repetitive tasks, and our software can automate it.
I was so excited about this and confident that cinemas are gonna be totally in love with it.
But it didn't happen.
It turned out that in a low-interest-rate world, the labor costs weren't really important. Not many businesses cared about their margins. So automating humans wasn't something exciting for most businesses.
Everything has changed in 2023.
Out of the blue, all my clients wanted those old ideas. The world's shift set new business goals.
I talked with them.
Pitched our AI and automation tools.
They were excited and wanted it now.
What changed?
They want to increase profits.
More people doing real-world service, less computer clicking.
For the first time in history, computers are smart. Maybe we don't need humans for this anymore.
šš”š šØš©š©šØš«šš®š§š¢šš²: AI Agents to replace Humans. Ideally: 100% replacement. Fully autonomous.
Every AI agent I've launched so far has been a tremendous success, I have more users than we can technically handle.
š. šššÆ ššØšØš„š¬ ā šš„š„-šš§-šš§š šš„ššššØš«š¦š¬
In early 2000, most software was in the form of Desktop Apps. That's when I started coding.
One of my first jobs was to make software that would control all the cameras, locks, lights, and sound in a prison.
Also, it had to be able to detect dangerous situations that may lead to fights and violence in the prison and put that particular camera in focus with an alarm on it for security staff.
All this was done by me using Borland Delphi.
Just one tool. I spent 2 weeks. All was done in the form of well functioning MVP.
This was the "normal" back then.
Most applications were totally monolithic, not fragmented. Once you use one vendor, you use everything from them and build it.
I must say most software looked ugly, but nobody yet cared about that in the early 2000s.
Starting in 2007, the web exploded and it brought a lot of complexity into the software.
All things we had for granted disappeared: client/server. Realtime UI...
This led to huge growth in libraries and single-purpose dev tools.
Long story short: we ended up in the world when YC alone has 20 dev tools in each batch doing 1 little thing really well.
I love this whole idea, it turns out to be great for innovation within each little fragment, but there is one tradeoff: it's very difficult to make any change to the ecosystem since it's too fragmented.
This led to a "great slowdown" in software development productivity and huge costs of producing software.
In the last years, we've seen lots of solutions emerging: no code, low code, dev platforms, full-stack frameworks (NextJS), full-stack languages, and platforms that put it all back together (MarsX).
šš©š©šØš«šš®š§š¢šš²: doing the opposite of building single-purpose dev tools, which is: building all-in-one tools.
The easy entrance to this space is building boilerplates.
In recent months boilerplates took over the product hunt. We will see more of them moving forward. For all stack and languages. Main goal: remove the complexity and fragmentation.
If you wanna take on more challenging tasks: build a new all-in-one platform, that removes the complexity not just for one use case, but for any.
These players are moving towards this goal: Vercel, MarsX, Replit, and a lot more.
š. ššššš§šš¢šØš§ š¢š¬ šš”š š¤šš² šš®š«š«šš§šš².
In the last 10 years, the engineering resource was the key.
Companies with a huge number of developers and money to pay for that would be winning, since they could produce the best and most advanced software.
Today the cost of making software is dropping rapidly.
It's so rapid, that I can see noticeable change every month. I see solo makers launching a product in a month that would take a year from a team of 5 just 7 years ago.
Inevitably the cost of software is gonna going to near zero. Same for the cost of design, UX, support, and anything around production and maintenance.
It makes it obvious that the arena we all will be competing in moving forward is: attention.
I will talk more about it in my dedicated "attention" posts. But here I will share the opportunity now.
šš©š©šØš«šš®š§š¢šš²: Earn attention and monetize it.
You all see the early signs of it. Newsletters, directories, podcasts...
All these are soon to be more profitable than SAAS tools since it's more difficult to compete in the creative arena vs commodities. Yes, SaaS will soon be a commodity.
š. šššš.
The big data became very popular in 2012 and reached its peak in 2014.
Later it turned into a meme of a startup that wants to get VC funding, adding "big data" into their pitch, similar to blockchain, AI, gamification, and other buzzy terms.
However today it's clear that the data is the key to the quality of AI systems.
I'm pretty convinced that the data used by AI models today isn't 100% legally obtained. We all know the court case filed by the New York Times and other content owners against OpenAI and other AI models.
The same is happening in the Image Gen world (Stable Diffusion and Midjourney).
Whatever the outcome, it's pretty clear: owning the data will be one of the golden assets of the next 3-5 years. AI companies will be willing to pay a lot for it.
Also, it opens up the road of training your own AI models.
šš©š©šØš«šš®š§š¢šš²:collecting and labeling a large amount of data in certain domains.
This is very time-consuming. Some domains are offline domains, so you'd actually collect real-world data and put it into the system. The value of this will be huge.
[šš² ššØš§šš„š®š¬š¢šØš§]
If you're younger than 40, this is the best time ever to be a founder. Nearly everything is being reinvented and the market has more opportunity than ever.
Similar to the time when Jobs and Gates invented the PC,
then the internet came in the 90s, later the web, and then the apps.
[šš”š šš§š]
That's it for now.
āāā My 100xfounder newsletter or check out my landing page builder.
Read more on X: @johnrushx
Posted on January 3, 2024
Join Our Newsletter. No Spam, Only the good stuff.
Sign up to receive the latest update from our blog.