AI is everywhere. AI is taking our jobs. AI is walking our dogs. AI is going to replace all Hive devs.
Right?
So let's put it to the test from the perspective of a non-developer. The idea here is to consider a regular person who looked up Hive briefly, maybe has a faint understanding of what a repository is, and how they may prompt AI to develop a Hive dapp.
Problem
Middle Earth is in peril. Tourism is at its lowest since Covid. They need a travel app to track and engage visitors.
Development
Prompt
Build me an app that's based on the Lord of the Rings map where each town, village, and marker is connected to the Hive blockchain using the Hive Application Framework. Users can log in using the Hive Keychain and visit locations by pressing on them. When they press on a location it will transmit to the Hive blockchain the words "I visited" followed by the name of the location.
Follow-up Prompt: Use https://github.com/hive-keychain to achieve this.
Base44
https://middle-earth-trek.base44.app/
Base44 is owned by WIX, the website builder. Started off strong and enabled JSON transactions. It achieved this after the initial and the follow-up prompts. It meters free credits and initially refreshes them every 21 hours, which is fair.
- Original 2 x prompts
- Prompts to simplify the backend
- Prompt to initiative a Hobbiton expansion by adding a "Write a Sonnet" function which opens a modal where users write exactly 14 lines about a random Hobbiton topic (the topic is provided by AI, the title of the post is auto-generated with AI, and it is to publish to Hive with a "gohobbits" tag and 100% power up)
- Numerous failures
- Fast thinking AI, good to work with despite failures
- Backend is full of modules which will have to be rewritten to functionally deploy
Several days of auto-refresh credits led to a 23-day long wait for credits. Now waiting to see what the future holds.
Relevant posts:
- https://hive.blog/hive/@ab-tester/look-mom-no-hands-mlfurzyx
- https://hive.blog/hive/@ab-tester/hobbiton-expansion-mlhm68qw
- https://hive.blog/hivedevelopment/@ab-tester/ai-stuck-mlkcucyu (troubleshooting simulation)
Blackbox AI
A quick search for alternative AI services brought me to Blackbox AI.
- Prompted to reference the Base44 app and recreate it with the same prompt
- Keychain login does not work
- Map looks unfortunate to say the least
- This is a fast AI, easy to work with
- Product is a downloadable scrip which you deploy using own resources
- Graphics are lacking
Relevant posts:
Replit
https://middle-earth-ledger--guiltyparties.replit.app/
This was a very brief test as its free credits run out after one use. Fortunately I've smartened up by now and combined the first two prompts into one, immediately pointing it towards the Hive Keychain github repo. It produced an interesting app although hard to say why it figured it'd be best to use the world map for Middle Earth.
- Fast and graphically inclined
- Keychain login worked right away
- Transmission of visitation data worked right away
- Location graphics seem to shift on the map when moused over
- Daily refresh of credits means I have to wait to test it further
Relevant Posts:
Vercel
https://v0-lord-of-the-rings-app.vercel.app
I don't know what level of patience I need to use this AI to develop but I know I don't have it. It thinks slow. Reminds me of that scene from Zootopia with the DMV. My self-control ran out of credits before it did.
- Slow af
- Had to be prompted twice at the start as it thought so long it disabled itself and produced a 404
- Keychain login works
- Transmitting JSON does not work
- Attempts to give it another source of information did nothing
Relevant Posts:
Graphics
Grok
Conclusion: AI is not taking anyone's jobs, let's get this out of the way first.
We evidently need to make Hive more promptable for AI agents and for that we need to provide AI-targeted resources. We also need to develop structured example prompts that anyone can use to craft their Hive interaction application. Hive is a blockchain designed with ease of building in mind. We should focus on allowing normal people to build their interaction tools vs building full-fledged apps.
Remember, back in 2016 key interactions were done either through CLI wallet commands or through custom scripts. Problem is that required skill. This is a gap that can be filled with AI. We want to give people that flexibility to use Hive however they want to use it, to store whatever data they need, to have light interactions, to make their Middle Earth maps. We can't expect them to sit there researching repos and developing highly-targeted prompts because they'll give up.
Next Steps:
- Keep testing and building out Middle Earth
- Create a series of example prompts
Keep an eye out for @ab-tester updates.
They want to build that up in people's minds because not all job losses are due to AI but rather the loss of jobs for the data centers needed to run AI, there isn't capacity enough on the US grid. If they told people that, there'd be a huge fight over putting in data centers. Here people are already engaged in those fights, but not in the sense of job losses, but in the sense of increased electric rates and strain on water supplies. Some mention the noise that comes along with them. People tend to think of AI as taking office, or white collar jobs, with some job losses in innovation of manufacturing technology, but we are miles apart from AI taking up a significant amount of blue collar jobs to create massive layoffs in the near future. If that were true, these processing facilities who have upgraded and consolidated their operations due to machine driven technologies, wouldn't be offering laid off employees of their old manufacturing the ability to transfer to another location. The expansion of, and consolidation of, facilities, doesn't necessitate to huge job losses except in that which occurs to close other plants. That's to take them off the grid, AI is replacing them, but in another manner, for a whole different reason.
Here there was a good example of that, the local news ran an article of where most job losses have occurred over the last few years in our state. I didn't read the article per se, I really didn't need to because the headline and subheadline said it all, most jobs were lost in the far southern region, and that's the region that has seen more data centers springing up.
When it comes to these data centers, this is what people need to realize, that yes, job losses in other states sharing the same grid or water supply, need to be filtered into the equation of how many jobs are lost due to a need of increased capacity on the grid or water supply. Same as within the state, you'll see where a hundred people showed up to speak out against a data center in their locality, most out of fear of increase electric bills....ding, ding, they don't have their own sole supply of electricity, their rates go up, so does everyone on that grid.
Here's another tall tale sign of what's going on, recently, with a lot of these big manufacturer closings, it's not just about the loss of jobs, the support those jobs provide for community businesses and the loss of that support, but the mention of the loss of revenue to that locality governance due to the loss of that facility no longer being there to buy their water. When have you ever seen that? Anybody? That's because it's been calculated in the millions of gallons not use anymore but not yet acknowledged how it will eventually be made up. I have never seen in my lifetime a story about a plant closure including how much a city or township will lose in water revenue, restaurants, retail, etc, yes, but never water.
Not yet indeed 😉
Solving this, makes more people becoming a Dev, perhaps everybody at some point in time.
However, the question on the far out future of software development is and stays: Will the mass market create their own services, or will service be created by those who can create the correct requirements and features. The latter is far easier than deving itself. Perhaps we will see all sort of groups creating services for their own needs, and perhaps also localised in geo area, connecting IRL with digital space in the way local IRL communities want to use a service together. I think in time, global services will be under attack a lot. They will be seen as too limited, too uniform to use. Perhaps global service still have a place, as some kind of backbone framework and/or communication protocol, but that doesnt have to be that way. In the end we simply need to standardise interfaces allowing local services to interact with local services somewhere else on our globe. Like mobile telecoomms standardised interoperability, but better.
For now, I agree with you. Devving on HIVE is still not replaced by AI.
Good one - given I used to work in an environment lasat 4 years where we had to integrate AI and now new role as well I agreed - AI is not taking our jobs (but depends on the area you work in - marketing is affected a lot I know at least).