Over the past few years I’ve actually gotten more optimistic about the future of humanity.
- Seen how fast we’re able to move when enough interest and resources are on board in important moments. Especially in open-source.
- Noticed the upsides and downsides of EU-cookie-banner-causing styles of regulation.
Online however, I keep seeing viral arguments appear that turn deeply doomer, like these two:
- Shut down or highly limit all large scale LLMs over water use concerns.
- AI is a slop machine without real value. Cool new innovations in the space are useless and are burning the planet.
- One example about using Genie 3 to “walk around” still photos or art said, “If you can walk around in photos then the whole reason for photography’s existence is laid moot.” Like, seriously? There are so many useful cases and loads of people pile on agreeing that it’s worthless
And then in person, I started hearing them too. Then I saw this post, which made it make more sense:
I consume a lot of information. A lot of podcasts. A lot of articles. A lot of research.
And I honestly can’t believe how many people are just parrots of information
Like most, I’m also a parrot of a lot of things of course, since it can drive good discussion and is what’s most comfortable a lot of the time. But when it comes to these topics that people get really worked up about, I want to weigh all the major options we have.
It’s easy to be the doomer because if you’ve only had useless experiences with new technology, it seems overall like it’s useless. But it’s also easy to be the naive techno-optimist, which I fall victim to, partially because of “magical” feeling moments and great marketing. But also because I’ve seen firsthand just how incredible some people are and not just incredible, but passionate about solving most the doomer concerns.
With our current trajectories in science, I feel overall pretty excited and don’t see how it’s worth it to be constantly on the doomer or deceleration side. I see a lot to look forward to, even related to solving big doomer arguments above:
- Data centers getting way more efficient over the past 10 years and even on renewables. Now with the huge AI push, there’s a huge surge in electricity usage driving for more nuclear energy.
- Chip architecture changes like with Cerebras give mind boggling AI inference speed results that could improve energy efficiency and human output efficiency
- Hypothetical (sorry): can this accelerate research across hundreds of studies to build out things like cancer treatments, improved clean energy data centers, or even fusion advancements faster?
Just from my own experiences in the past year alone, I can’t agree on much with the hardcore online tech-dystopia doomers. I’m using local models for translation, learning, guidance, and programming. Gotten myself into many rabbit holes and accidentally into like 3 new major hobbies, including going outside a lot more. I’m not sure how that’s useless, but get told every once in a while that somehow my experience is invalid?
A popular tech take currently is that, “even if our AI models never get better from here, there’ll be a lifetime of things to build out.” I agree, since the plumbing portion of the work I’ve done over the past year has been grueling, and continues to surprise me. But this plumbing excites me, since even if everything stopped right now, there’s so much to explore. And this is even just on my computer, without having to “burn the planet” :)
It’s an uncertain and unique time, so I totally understand that it can make people turn to pessimism. For example, the pace of AI scaleup compared to the creation of nuclear powerplants or scaling up desalinization, aren’t really aligning. I just feel like that could be the core of a pessimistic argument. I don’t (in online discourse at least) frequently see action towards specifically ramping up those areas, more so just anger towards businesses taking advantage of free money. I feel like that subtle directional change in energy could make a big impact.
Comments
Loading comments...