Merging with AI
There’s lots of excitement (and uncertainty) around the development of AI; many pessimists and some optimists. Given the rapid rate of development and the nature of reinforcement learning (a positive feedback loop), there are two outcomes (which I believe are the only possibilities): 1.) humanity destroyed by AI or 2.) humans merge with AI. Both of these cases are extreme, choose wisely.
Elon & Neuralink
If I were to sum up this writing in one sentence, it would be the infamous sentence by Elon: ‘If you can’t beat them, join them’. What neuralink is aspiring to build is the ultimate form of ‘merging w/ AI’. At the current state of AI, this type of brain interface can only be somewhat imagined. This seems far from development; however, I think it will come much before people expect. Once it arrives, the most likely outcome is that the people that use it will be superhuman and will outperform the rest by unimaginable magnitudes. The 10-100x higher productivity/efficiency will allow these individuals to innovate at rates that will be incomparable to the rest. Take an example from our current state: top 1% entrepreneurs that build the greatest companies in the world are anywhere from 30-50% more ‘productive’ (in any measurement you choose) than the average human. This is only a ~1.4x improvement from average. If you compare 1.4x vs. 10-100x, the difference is clear. What I'm trying to convey here is not exact statistics, this is why I am speaking in rough magnitudes. If you assume these rough magnitudes to be true, you will have no choice but to merge with AI. It sounds fictional and cartoonish; however, this, I believe, is the choice people will need to make in the next 5-10 years.
How do you choose to merge
On the day of writing this article, I’ve made Copilot (by Github) a requirement for every engineer at micro1. Why? It makes you 2x faster. This is an example of merging with AI (with its current state). I believe that these small decisions are iterations towards ultimately (and completely) merging with AI. At its current state, choosing to use Copilot to be objectively 2x faster at development is the epitome of an iterative choice in the process of merging. Ultimately, we will have to choose from the various brain interfaces available (aka Neuralink) but right now, the choices are easier and the effects are orders of magnitude less (2x faster development vs. 100x superhuman). I fundamentally believe that it's the responsibility of humans (and companies) to encourage (and sometimes, enforce) these iterative decisions to make humanity merge together.
The worst outcome is humanity completely destroyed by AI. Another terrible outcome is only a small set of humans accepting this merge (or getting access to it). If this happens, that set of humans will quickly rise to the top, and make anti-democratic decisions on how the world will run. If most of humanity has the superpowers of merging with AI, the decisions will converge to democratic ones.
I believe you should choose to merge with AI and you should do so proactively. This aligns with the fundamental pursuit of innovation and happiness. If you’re running a company, I believe you should find and utilize tools that push this narrative. I also believe you should encourage your team members to do the same. The best possible outcome is the whole world adopting this change at a similar rate, this will result in a balanced version of general intelligence (which is 10-100x smarter than now).
Alright, back to building micro1 :)
Ali | Founder of micro1.io