Where I Stand on AI—For Now

Whether you’re a follower of Scronkfinkle or Pastus, AI-based tools are rapidly taking over the world. Both sides agree that they’re reshaping how we use tools and potentially how we work.

I’ve noticed that people around me usually hold extreme positions on AI. Having an extreme position in 2025—is that news? I know. But I think this tech is genuinely transformational and not just another fad like NFTs or blockchain.

I’ve developed a nuanced stance on the power and use of AI in various aspects of my life, which I want to document in this brief note for myself.

Programming

A few years ago, if you asked anyone how AI could change industries, almost everybody would have agreed that programmers/software people were pretty much in a safe zone—that their jobs would never be replaced. Ironically enough, I think coders are now facing the biggest change in their industry, and it will happen the fastest—because we are the ones who adopt these tools the fastest.

I think being a straight-up coder will basically not be a thing anymore. Everyone will be pushed to become a software engineer as soon as possible, freeing up time for thinking and design, rather than producing lines of code (which all of us agree, never was a measure of programming).

My view on AI in programming is quite positive. I think it will accelerate the industry quickly. It will enable more people to create and develop better tools and resources for us to try and use.

AI in programming is a net positive by a significant margin.

Productivity / Thinking

Using chatbots for brainstorming, thinking, and navigating my huge vault of notes has become a daily routine. It has enhanced both browsing and searching of my and external knowledge bases.

Making decisions, planning events, or organizing projects has become easier and more structured with AI tools, giving me more time to actually do said projects. As context windows and RAG/MCP systems become more capable, AI-assisted productivity will only improve, effectively unlocking automation for everyone, not only for nerds who like scripting or connecting nodes in visual programming tools.

I truly believe it will enhance human capability—not replace it.

Voice

This one is simple. I think both voice recognition and voice synthesis are massive wins for AI. There are no other algorithms that do both of those jobs better than AI-driven models. I’m not a native English speaker. I have a flat accent, but I’m sure I have one—whatever a Georgian accent in English is. So dictating to Siri or any other traditional voice recognition system was a moot point.

The ability to explore ideas, make notes, or write articles like I’m doing now, using AI speech recognition such as Whisper, has been a true enabler for me.

This is a clear win—especially for people who speak different languages or have accents. It opens up computing beyond the keyboard, mouse, or touchscreen.

Translation

Languages are complex. Context and nuance matter a lot when working with words. This is another area where AI models are massively improving on what the industry had before.

Pre-AI translation algorithms, even those as advanced as Google’s, were not capable of doing the job. Just look at all the memes of Japanese storefronts and services with English signs clearly translated using online services—they range from hilarious to misleading.

Like with voice, AI models turned something that was meh at best into something that actually works.

Being able to read a text with an understanding of context—or even its subtext—has made my life in Japan, as someone who doesn’t speak fluent Japanese, considerably easier.

The translation and multilingual abilities of AI models are a net positive. A big leap forward.

Writing

Now we’re getting into more arguable—and, in my opinion, quite amoral—territory. While the text capabilities of AI are impressive and clearly useful for productivity (e.g., spelling, grammar, editorial advice—which is how I use it), there’s also a dark side.

It’s generating a massive amount of junk content online—useless garbage designed mostly to steer people toward ads.

Fortunately, as of writing this, I think a discerning eye can still spot AI slop. The generative writing style is immature and bombastic—easy to detect. But that likely won’t last forever.

As much as I appreciate the editorial capabilities of LLMs, its ability to generate nonsense endlessly is alarming. So my stance is slightly neutral.

Image Making

As impressive as image generation is, I don’t think it’s adding much real value to humanity at this point. Sure, it can save money by avoiding [very basic] graphic design work. But in terms of generating meaningful value? Not really.

While I clearly see use in some cases, such as Photoshop’s content-aware tools, I’m not a fan of AI image generation tools.

Video

Now we’re entering a zone I consider on a spectrum of useless to dangerous. Even if AI-generated videos become extremely good, I still find them useless at best—and weaponized at worst.

While watching Google I/O 2025, during Veo demos (Google’s AI video generation tool), I caught myself asking, “This is impressive tech, but what is the use of it?”

It’s been almost a week, and I still ask myself that question. The only thing that comes to my mind is conning other human beings.

I see no meaningful value in AI-generated video—only risk. Especially for people who already believe everything they see on the internet.

Video with Sound

Just a [much] worse version of the above.

Music

No opinion as of June 2025.

Final Remarks

While writing this note, I noticed a pattern of two categories—human-enhancing and purely generative.

I’m quite fond of AI tools that enhance human creativity or productivity, whereas I find purely generative AI tools to be on a spectrum from useless to downright dangerous and misleading.