Imagine this you used to need a library for common things in your software project. Even if you just need one function but because it was easier to just import a library that would have been the standard practice. But now the AI will just go “I can just implement that thing you need in 10 lines”. You used to need things like react native or flutter if you wanted to build cross platform apps. Now not anymore you just need to tell the LLM and it does it in both, and you get better results too.

In society we also had all these layers of abstractions and hierarchies that we used to need but will become more and more irrelevant collapsing the hierarchy.

There is a saying “as above so below, as below so above” I think this applies here. It will propagate all through our social construct, software, society.

I think this is where the role of humans will move towards in the future. Things like accountability can't really be outsourced to AI.

This relates to solving the alignment problem: if you give the AI a specific goal and it has to plan sub-goals, how can you be sure the sub-goals align with your interests.

For tasks in an isloated environment, this doesn't really become an issue. You can give a sandbox the least privileges it needs to do the job you want it to do.

It becomes a problem if you give it open-ended access to systems that connect to the real world and which can have real consequences. There have been stories about openclaw deleting someones email inbox because it though that was what the owner wanted. AI taking control of public wikis to coordinate, which was recently posted is another one.

For software concerns with potential real catastrophic consequences are security, durability, availability. I think for future software systems these are the concerns where you need to limit the AI in a way that it cannot circumvent guarantees that you give around these concerns.

Minor nit:

> AI doesn’t change the question: it creates new choices and moves the thresholds.

This cause an adverse reaction when I read the post. This was probably not written by an LLM since the rest of the article doesn't look like it, but maybe in the future we need to all be more concious about leaving LLM-tells out of our writing.

Giving people AI is just like giving people Google Wave (https://en.wikipedia.org/wiki/Google_Wave), which can do pretty much anything collaboratively, ended up doing nothing.

There are a lot of redundancies within the tools and the pricing is such that one cannot do much about it, generally some companies pay for the brand, and some tools like SAP are integral to companies of certain size. A sufficiently integrated AI tool that learns the workflows might actually be able to find optimisations here as well, but definitely auditability, testing and other concerns will remain and this is what might become the USP of SAAS providers.

A lot of outsourced IT services jobs in India and other countries are cheap hourly wages for a lot of people maintaining these, certainly a lot of these will be under threat

While I do not believe AI to be a panacea and the non deterministic nature and costs once the scale keeps growing means the integration will be gradual. I am still excited for it to define what an organisation is and what do a lot of people actually do especially in fields like accounting etc. where repetitive work is billed at quite high rates.

Law etc. is a field where the gatekeepers might hold on much longer by adding more ridiculous rules and logic. Ultimately humans have decided what is constitutional and what is legal and subjective rules are what maintain human power.

I feel the right model is smart domain experts of humans making strong and useful harnesses that help AI be effective with the workflows of the organisation, however this might be the biggest fear of middle managers who will never let it happen easily

However, things are changing so rapidly that I can see that starting to change as well. But it would take much better learning efficiency to understand unknown domains, 100% computer use reliability etc. I suspect we’ll see this by the end of the decade.

It is not imagination, it is the undeniable truth. The quality of the tool is proportional to the competency of the builder, it always has been that way, AI makes the building process shorter by a magnitude

It will raise the level of personal agency + stable family resources required to avoid crises and homelessness.

Society will increasingly organize around preventing class warfare.

I really wish those farcical statements would stop (I know it's not really made by the author itself).

I'm a software engineer, my partner isn't. She used AI to vibe-code scripts to help her analyze large Excel files. It worked, except the script only looked at the first tab of the Excel files, not all of the tab, which could have had disastrous effects if I didn't read the code and realize it was wrong.

Similarly, her team went from using Excel to track project proposals, to vibe-coding a website that's deployed on CloudFront and AWS Lambda. They don't understand any of this technology, have no clue what they are doing, and, more importantly, don't realize that they have sunk tens of hours, if not hundreds, in developing a portal that has absolutely no advantage over the Excel file they used to use. They also don't realize that if I wasn't there as an experienced software engineer fixing the stuff that's broken when my partner asks me for help, their website just wouldn't run at all.

No, non-engineers cannot make their own tools.

I was very much in the doomer category last year, but less so now. For someone working in Accounting/Finance/Revenue roles, who has always had a keen interest on tech and software development (but never dedicated enough time to really learn a language to build something of my own), I had many ideas on tools that could be useful in my job - just never had the skill to build it.

Now that I have AI, I'm experimenting with building some of these solutions, but I must say: software development is hard. Yes, it would be a lot easier if I had others to work with, especially for things like UI/UX, understanding some of the more complex backend decisions etc. I imagine if I find it hard, people that haven't really been following tech trends would find it even harder.

Yes, you could get away with vibe coding a tool that only you would use, but once it goes beyond that, and it needs to be shared with colleagues, or if the aspiration is to market it, there are so many decisions one needs to make to be at least close to world class quality.

Things may change as more advanced models come about, but at this stage, I still think unless you're a seasoned developer with good design taste and UX experience, you need a team of people with various skills to build something that you hope to GTM with.

It's fun, though. I've created a few tools internally that has saved me so much time and developed functionality I would never have even entertained the idea of pre-agentic coding. I hope my career moves toward building solutions in this domain.

> maybe in the future we need to all be more concious about leaving LLM-tells out of our writing.

I expect the gruesome writing style of LLMs to be fixed relatively soon, seeing how it can already be prevented with just a few lines of stylistic instructions.

I just wonder: how long will it take us then, after the fix, to not be thrown off every time we see some unnecessary catchy or contrastive phrasing?

Being traumatized by poor writing was not on my bingo card for 21st century technological progress ...

Accountability will not come before AIs achieve legal personhood, and that probably will not happen in my lifetime. And if it does in some jurisdictions (which I am not betting money on, this is a full full AGI scenario after multiple more philosophical goalposts move first), I will completely bet it will not be a global recognition.

Like, you need a guy who understands Linux and be accountable for Linux working well in your company. But if Linux does a good job that doesn’t leak, you don’t need a Linux guy.