shahbaby

I think right now AI is helping the attackers more than it's helping the defenders.

That part is probably obvious.

But in the long term, I suspect this will get worse.

Attackers will still need to think critically to be a few steps ahead of the defenders.

But the defenders will increasingly outsource their thinking to AI as they are incentived if not outright forced to dumb themselves down into interchangeable cogs.

I suspect this is also Astroturfed.

Real people don't usually think like this.

Too bad, I'll contribute less to this site now that AI infestation seems out of control.

This post is Astroturfed AI fear.

I'm guessing the AI oligarchs want fewer developers in the future so companies will become increasingly dependent on their tools.

A similar thing already exists today but in some ways it's worse because the economic harm can come from anyone with a grudge against you.

Introducing, Google Maps Review in 2026.

Anyone can say almost any type of slander about your business without even being a customer. If you report it to Google, the burden of proof is on you to show that they violated a policy.

How can you prove that a review came from a personal grudge vs a genuine review of the business? Sure you can say that you have no record of that reviewer, but how do you prove it? Google rejects nearly all such reports and appeals without giving any reason.

A Google Map "review" is more like a public messaging board where anyone can write whatever they want and as long as it appears loosely related to the business it'll be nearly impossible for the business owner to take down.

I think the best thing that business owners can do is take matters into their own hands by showing genuine reviews on their own site.

Seems like a lot of survivorship bias here.

I think intensity is a symptom of circumstance not a cause for them.

If you're a startup that's on the verge of transitioning from broke to huge success, I'd imagine everyone would be bringing a lot of intensity to push it over that line.

If you're an established company with a track record of responding to record profits with layoffs, then I don't think it's fair for you to expect your employees to share your enthusiasm.

Because LLMs still do not know how to manage context.

Proof: - Load up all your chats, across several different topics and see how even the best models will try to connect dots with lines that don't make sense.

As a side effect of the above, LLMs have trouble coordinating with other LLMs.

Humans are able to compress and manage context without much effort and each human brings their own unique set of experience and perspectives.

That's why humans scale in ways that LLMs cannot yet match.

Maybe by next year they'll crack this too but that's where we are today.

Might be an Apple UI designer.

If you can brand, design and deploy entire products on your own, try doing that all with AI.

See where the gaps are and from that you'll learn which areas you should focus on.

In other professions, workers protect their own. Doctors are the best example of this. They have regulatory agencies that ensure the supply of doctors is always kept artificially low hence why there always appears to be a shortage of doctor while at the same time med school has a very high barrier to entry, it's all by design.

Software devs are the polar opposite of this.

- We have the most gruelling interview process.

- We normalize working over time with no pay (oncall).

- We openly invite people from all walks of life to join this career. To suggest any sort of minimum standards is viewed as gatekeeping. Learn to code even became a meme at one point.

- We are supposed to sound like our work is our life and even outside of working hours we're busy with side projects.

I don't know of any other career in which the workers are so brutal to each other.

I hope the current AI revolution is a wake up call.

> And that skill is? Any concrete verifiable examples? That can be replicated, repeated by others or adopted by others?

I learned this weekend that 1 LLM working on its own with all the context is usually better than splitting the task up into sub parts and having multiple LLMs work in parallel.

The interesting part is that I had read about this previously but it wasn't until I had seen it on something I was working on did that principle become engrained.

I'm sure there are still ways to make this work. i.e You could specify how the context should be passed between agents.

Knowing when you need to do that, what works well and what doesn't, etc., is where skill comes into play.

I'm not sure if I agree with this but I find the downvotes at even the suggestion of some mechanisim to protect employment sad.

No other profession is as eager to make themselves obsolete as those who build software.

Maybe they think the world runs on meritocracy. Anyway looks like now they are seeing what happens when there are no gates at all and anyone can do what you do.

I've accepted that the "Old Internet", the one in which most content was clearly human generated (for better or for worse) is dead.

On the flip side, the AI infestation makes the New Internet, at least for me, less addictive.

After you've filtered down the resumes, if it's not financially worth it to put a human on the phone for the initial screen, that means you don't value the people you hire.

And if the cost to do that initial screen is almost zero, it also means that you may not even have a job opening and are just testing the waters.

No matter how you slice it, it's a bad look.