Why Steem stakeholders should embrace AI incubation

As far back as 2016, well before the current LLM hype, I already believed that Steem's technology represented a groundbreaking platform for AI incubation. At the time, I wrote,

If (big IF) steemit survives as a company, in 5 years or maybe 10, steemit's ecosystem will have better AI than Google or Facebook or any other closed company.

And here was my reasoning:

It's The Cathedral and The Bazaar again. Facebook and Google used to be bazaars, but now they're cathedrals. New bazaars must arise. Why will steemit be one of them?

• Selective pressure from flagging .
• Constantly growing and freely available training data on the block chain, and in live posts and comments. (Note: this blockchain is also available for harvesting by advertisers.)
• Large incentives from author and curation rewards.
• A planet-full of steemit bot developers will be more productive than the staffs that Google, facebook, or anyone else can afford to hire.
• Unexpected and surprisingly useful emergent behaviors will arise from complex bot interactions.

So, obviously, I was wrong in terms of timing. I still believe I had the direction right, though. I posted on the topic again, yesterday, in Steem as an AI incubator - again: The StoryChain.

Why do I continue to believe this?

Basically, everything I wrote above is still true (though incentive sizes are smaller now). And I had not yet come to appreciate these other aspects of the Steem chain that also point in the same direction.

1. Free transactions make micropayments feasible (see the 402 Foundation).
2. Fast transaction confirmation allows micropayments to scale.
3. AI is an ideal breeding ground for the use of micropayments.

So, in short, we have a blockchain with a unique rewards mechanism and underlying transaction economics that still seems like it should be ideally tuned for AI to be tremendously successful here.

OK, then why hasn't it happened?

I divide this answer into two epochs, 2016-2020, and from 2020 forward.

During the first epoch, the reason it didn't happen is that there was a dominant ring of early miners and SP stakeholders who were determined to prevent automation from gaining a foothold, and they were able to successfully strangle innovation through the use of downvotes and reward starvation. Most of them left the chain in 2020, though.

From 2020 and forward, it's a harder question to answer. Nobody has been 'policing' blockchain content effectively, so innovation probably could have happened, but it mostly didn't. My best guess is that this is due to three primary reasons:

1. Reputational damage that happened during all the drama of HF22 and HF23. It seems like Steem is finally on the cusp of recovering from all that damage, but it slowed everything down. And, it must be recognized, that overall market conditions for the STEEM token and other altcoins continue to be punishing for investors.
2. Uncertainty from Biden era SEC policies.
3. Steem's dominant investor for most/all of the second era doesn't seem to understand or appreciate the things that distinguish Steem from other blockchains and cryptocurrencies in his portfolio, especially the free transactions, the custom_json operations, and the programmable rewards system (programmable through the use of beneficiary settings).

What if Steem's designers got the primary use case wrong?

With AI incubation in mind as an alternative use case, I asked Brave Leo (Qwen 14B) the following question:

Can you name a product or commodity in common use that was first designed or harvested for one purpose, but wound up being wildly more successful for another purpose?

Leo responded with two examples: The sticky note, which was first designed as a lightweight adhesive to hold papers together; and the microwave oven (cavity magnetron), which was designed for use in RADAR during WWII.

I asked Leo to tell me more about the microwave, and it turns out that its potential for heating food was discovered by a radar technician who would regularly notice his candy bars melting in his pockets. 80 years later, the cavity magnetron is still used in radar, but that use is dwarfed by the device's use in food preparation.

So, we definitely have examples of products that were intended for one purpose but succeeded at another. In at least one case (the microwave), the original purpose is still being fulfilled, but a much more useful purpose was also discovered. And that brings us to the title question.

Why should Steem stakeholders embrace AI incubation?

Let's suppose that social media and AI incubation are to Steem what radar and microwaves are to the cavity magnetron. Why should Steem stakeholders embrace this?

1. Something has to change. We've been watching the token price collapse over most of Steem's 9+ year lifetime. What we're doing is not working - during either of Steem's two epochs. I, for one, am tired of putting in years of effort just to have it swallowed up in two hour market movements. I'm sure I'm not the only one. It is unacceptable for us to continue to tread water and hope that we eventually drift somewhere where we want to be. Treading water leads nowhere but exhaustion and drowning.
2. The time for AI is now. Market forces are favoring AI, and the industry at large is still learning how to profit from it. In fact, we may already be too late, but I hope not, since Steem has some unique and untapped attributes that it brings to the table.
3. If AI is properly plugged into the ecosystem, it will make the platform better for humans, too. The social media use case can continue to improve, even if we're sharing the technology with automatons. If AI incubation brings funds into the ecosystem, then applications with a human focus will also have better returns on investment.
4. The people who successfully strangled AI innovation during Steem's first 5 years are (presumably) now blocking the same sort of innovation on a competing chain. Let's leapfrog past them.

Conclusion

With fast and free transactions, Steem is an ideal platform for micropayments, which can be exploited by AI agents. With its programmable rewards system, it's an ideal platform for AI agents to get feedback about human preferences.

We should stop pretending that these things aren't true by typecasting Steem as merely a social media platform.

Investors who want to optimize their returns should be able to look at Steem as a welcoming place for AI-related investments, and Steem's upper tier stakeholders should aggressively start exploring ways to integrate AI incubation into their Steem ambitions.

Finally, the community should get to work on drafting and obtaining widespread agreement on some sort of netiquette for the use of AI on Steem. In a decentralized and permissionless ecosystem, an "acceptable use policy" is not possible, but we can still create community-driven guidelines.

I'll sum it up and close it out by repeating my opening claim from yesterday's post:

Steem needs to be a welcoming place for humans, but it doesn't need to be an exclusive place for humans. Innovative uses of AI can and should make Steem into a more appealing social and crypto platform.

Thoughts?