Next is writing code entirely. Whatever has the most robust documentation will be the next big thing, since AI agents will be writing the code for the programmers.
I'm concerned about the opposite possibility. All the effort will go into tweaking models and adjusting their training data until they generate acceptable output, and documentation for human consumption will be an afterthought.
> documentation for human consumption will be an afterthought.
The documentation just changed a layer upwards in the abstraction ladder. Instead of docs for the code that runs on the machines, we'll write docs for the programs we've built that can generate the code that runs on the machines.
This is a great project at CMU. Worked on it from the beginning as an undergrad.
It’s a very unique project: students have the ability to be involved in almost all of the roles of the project - from mentoring high school teachers to writing new course content and working on backend systems. There are 2 professors who oversee the project, and a handful of awesome full time staff to guide and manage the CMU students.
Reminds me of a passion project I started in high school that went completely viral and took on a life of its own. Wrote a small script for my friends to check their AP scores a few days early. Required high schoolers giving clear text access to their entire CollegeBoard account so I could log on and scrape their scores. Somehow it got posted to Reddit and from that year on, grew wildly. Got to almost 2 million students checking their score in its peak year. It was immensely fun while it lasted (ran for about 7 years) and honestly I miss the thrill of it. CollegeBoard now releases all scores on the same day so the site is pretty much useless now. Definitely always looking to chase the thrill of that score release day again though.
Congrats on a successful end to a fun high school project! Stories like this are always fun to read.
Can you elaborate on the approach? This sounds really interesting but I don't quite understand from reading your comment and https://earlyscores.com/about/
I think I remember paying some small amount of money (flat fee irrespective of # of years, IIRC) to get my scores quicker via a phone call in 2003,2004,2005. Perhaps I would've been better served by your EarlyScores.
The approach was fairly simple: access to the college board’s website was geo-IP restricted for about 5 days time. It would start with a small collection of states, and each day over the five days another group of states would get access to the site starting at ~8:00am EST. I would get a few AWS/GCP/DigitalOcean nodes in a DC that had an IP in a state releasing on the first day. Put a small JS script on the nodes that would use the username and password input from students to sign in to their Account and send back the scores. Basically just a proxy without the need for configuration.
> In 2014, with my first AP courses under my belt, I anxiously anticipated the release of my AP scores. What I realized at that time was that scores were rolled out by the College Board over a week’s time, and my AP scores would be accessible on one of the later dates. The need to see my scores on the first available date spurred me to create EarlyScores.com.
I remember that kind of thrill. For a while I ran a tool for Etsy before they had an API, circa 2007. What they did have was AMFPHP, powering their flash toys (treasury, etc). I used it to allow sellers to see their sales stats.
Even went to the Etsy office in Brooklyn at one point and had a chat about it. I think some of the team was a bit bemused that I'd essentially extracted a large amount of data. But they took forever to get to the point of having an actual API (and I was one of the early users of this as well).
Eventually it became unsustainable and I shut it down, but it sure was fun having people be passionate about using it and sharing it.
That’s not exactly true - as an Amazon affiliate you do see the exact items purchased under each of your specific tracking IDs, as well as the price it was purchased for, category and device group it was purchased using (desktop, tablet, mobile). This also includes any purchases the user makes in a 24 hour session of browsing after clicking your referral link to Amazon.
I’m unsure how many tracking IDs you can create in your account, and as far as I’m aware and can tell, you cannot pass specific UTM codes or other identifying information along with a click to Amazon that is passed back to you on the reporting side. Meaning, you could track users you send to Amazon, and where you’re sending them, and you can see outcomes, but Amazon only provides the tracking ID back to you as a reference (this ID is meant to be used on a site/channel wide level, but as I mentioned above could possibly be abused depending on how many you can create)
I got lucky, the subdevelopment I live in has been upgraded from classic Centurylink DSL to “Spectrum” fiber (1gbps fiber line to the house for $60/month, no contract and no equipment or setup fees. Growing up a few blocks away, it was a comcast subscription or a 1mbps DSL line(~10yrs ago).
It’s awesome when a competitor can enter the market.
In high school I made a site to help students view their AP scores early. It ballooned in popularity over the past 8 years. At it’s peak in 2020 and 2021 it was getting over 1.5 million students in 2 hours on the day of release. It was quite profitable for a side project I started in high school. Sadly I was never able to turn it into anything more than an early scores site on the day of score release, but now all AP scores come out at the same time and the site is fading away to being forgotten.
You must run earlyscores.com! That site was a life saver for me. Sadly anything aimed towards tech savvy high schoolers is probably going to be pretty difficult to monetize.
Not too difficult! Monetized with ads, the project was making a significant profit (>$15k per year; all of it came in on one day) - the difficult part was trying to do anything else with all of those users.
Tabler is a really awesome theme - sits right on top of bootstrap, and gives a very cohesive look for backends/admin interfaces. Also, the corresponding email template library is fantastic and only costs $30. Doesn’t seem like the project as a whole gets enough love.
One thing I haven’t seen mentioned is looking at Facebook Groups. The advertising can be weird because it is somewhat of a grey area, but most Facebook groups pages are run by companies or moderators who are looking for a little extra cash. They will add you to their group and let you post a certain amount of ads per day. You have to make the posts interesting and provide value to the group members, as more engagement will increase the reach of your post.
On a side project last summer, I was able to work with a company running education niche groups. To join 10 of their groups (total reach probably around 30k audience) was $500 per month billed on PayPal. Could make a post per day in all 10 groups, so 300 posts. The conversion rate was better than any other paid advertising we had tried and the targeting was much easier.
It’s easy to get started: find a few Facebook groups you might be interested in joining, go to about this group, add the admin as a friend/send a message request, and ask about advertising opportunities. Be sure to show them your project, and why you think it would be beneficial to their audience, that will help your case!
Also - have you thought about going to teachers with your product (assuming it’s compass letters from your post history), getting them on a free subscription, and hoping the kids go home and tell their parents they want to get the letters too?
Anecdotal evidence - but I feel like this could be completely valid.
One of my biggest side projects for many years was a student tool centered around test scores. It was a niche use case with a huge amount of students using the tool on one day per year (1m+). There was exactly one competitor. I had a better domain but a much worse site in terms of design, speed etc. We were nearly the same in traffic, until I decided to monetize the site with a lot of Google ads. Immediately, Google shot my site up in the rankings, and actually seemed to penalize the competitive site. My traffic went up 10x and the competitive site remained flat. This happened for 3 years, then the niche use of both of our tools was “patched”.
When an NFT project proposes a value share and airdrops amongst token owners, and also has a royalty fee paid on every transfer [1], I think the “up line” and MLM aspect come more into play. There is value for you to inflate the price of the NFT as it benefits you and the other owners directly as sale prices get higher.
