Ask HN: Software Engineers to follow who have a healthy skepticism of AI

CS PhD here. I've been involved in deep learning for many years now, and generative AI for the last couple of years. I saw some other major revolutions in my life starting with the Internet/Web explosion and the Smartphone explosion. You can say cloud and social networking were significant events but I would not put them in the same class as the first two. I would put the current embodiment of VR and crypto on an even lower rung of the "society impact" ladder. My personal opinion is that Gen AI tech (particularly LLMs but perhaps also the related stuff - diffusion, MLLMs) are more like Internet and Smartphone as game changers for society. Now, for the Internet boom, there was a ton of hype. I don't recall smartphones being as frothy but maybe I am forgetting the froth that was Angry Birds. So while I agree there is way too much hype at the moment, I think this is going to be big.

I have found it nice in the coding area to bang out snippets or auto complete the pattern for things like data structures and flow control. Sometimes even debugging but there is a large amount of oversight to massage the code to correctness. It saves me time on the meneal things but the higher level comprehension is nowhere near there yet. As far as getting through documentation, summarizing, and interrogating, it can be much better, especially if you put all the documents in the chat scope so it can use it.

My take is that it won't replace us but our roles will evolve to leverage these tools more and more to achieve human ends. There are so many considerations to take into account like production stance, fallback, data integrity, mitigation, privacy, security. I think we are a ways from that. When AI can start solving production issues post mortem and requirements implementation, then I would say we have hit a milestone. With that said, new engineers should still need to understand the fundamentals of CS, systems and how to properly use a language.

My main gripe is this push of AI in everything without fully understanding the value proposition and if it would be useful in the given case. I mostly ascribe this to leadership who don't really understand it and think it will just solve all of their problems and vastly increase productivity. What's worse is when it gets passed down the chain from C-suite and ends up being a hot mess by the time you get a project because there is no true direction of how it will apply to your product. It takes good leadership and identifying genuine usefulness for certain products.

I can't help you, but I'd also be interested.

The hype cycle around AI feels unlike anything I've ever experienced. Like yes, obviously it's a useful tech, but it won't become a literal god-like figure as some of our industry leadership is convinced.

I'm not even convinced it'll become reliable enough to replace humans for most tasks. I'd love to find some fellow skeptics that still appreciate and report about the actual advantages of AI as a tool, not as a religious movement.

I guess “thoughtful” and “well-balanced” are subjective, but here’s some content and a few “influencers” that I personally find more realistic and balanced:

- Simon Wardley (of Wardley maps fame) on Software Engineering vs Vibe Coding: https://www.linkedin.com/posts/simonwardley_x-why-are-you-so...

- Gergely Orosz from https://www.pragmaticengineer.com/. I find myself agreeing with takes like this one: https://twitter.com/GergelyOrosz/status/1912135400480526366

- Maybe a bit too optimistic, but I agree with the overall picture presented in https://sourcegraph.com/blog/revenge-of-the-junior-developer

I don’t do social media so I have nothing for you to follow.

AI is just the current tip of the spear. AI is most appealing to people looking for ever more convenient social exchanges, social gravities. While this is somewhat apparent in reality it is intentionally the focus around marketing of AI products and services.

There appears to be a clear and growing trajectory towards “tell me what I want to hear” in combination with “I’m not qualified to participate but can with just a bit of help” mentality. This thinking is driven by personality irrespective of education or intelligence, but those appear to be strong indicators as well. The personality types in question tend to be high agreeability and high neuroticism, which results in preferences towards social gravities versus individualism/originality. Historically this mode of thinking was considered to be more feminine but over the last few decades there been has a gradual erosion of masculine social identity in the west resulting in males adopting this more traditionally feminine social conduct.

https://antirez.com/news/148 (Redis creator)

I think he's doing an AI YouTube channel. He was quite rational with all the nosql hype back in the day, and seems to still be quite rational.

I will never use AI to any degree. Human minds are more than complex machines. We are creative in a way that can only be replicated by making more humans. Only a long enough time line will vindicate me.

That’s because human minds tend to do things like performative behaviors to mend fragile egos. Also things like hunting, pooping, eating, and lots of other things that im not sure that AI should concern itself with. But congratulations on your complex human mind.

You will never use AI that you know of..

I will also never willingly use slave labor that I know of, yet I have no idea where any of the parts of my computer or phone came from. We do our best.

Within the next few years, you will be using AI knowingly and constantly if you use modern computers or the internet.

I wonder if this is what it was like for people when they learned that the earth revolved around the sun.

I’m not a software engineer but I’m a data scientist and I’m pretty skeptical of it: https://open.substack.com/pub/tejo/p/why-ai-is-still-dumb-an...

I don't need to take a single stance on it.

Ai do 80 percentage of my daily job.