Daniel Miessler ⚙️

The further away a topic is from your expertise, the smarter an AI will sound.

This is a blind spot that not enough people are aware of or take seriously.

The best possible way to test the expertise of an AI system is to have it perform analysis or work in an area where you're a pinnacle expert.

Stepping even a few degrees off from your center-mass expertise will make it sound smarter. And the further you go, the smarter it will sound.

I recommend you have a series of deep questions within your area of expertise, where you know the arguments backwards and forwards.

Using those as part of your core benchmarks for testing new AI will be a lot more useful than the feeling of throwing random tasks.

Anyone with sufficient language skills can sound smart in areas you know little about.

Let me try to explain these AI layoffs.

The issue is the vast difference in quality of employee between the top 10% and everyone else. AI can’t come anywhere close (today) to replacing the top performers at a big company.

But they’re spending millions a year on tens of thousands of employees in the bottom 75% of the quality scale. And AI is now getting good enough to replace THEM. Not directly, yet, but by having top performers do more of their work with AI.

So it’s not that AI is better than top performers. It’s not.

The issue is 90% of employees aren’t in the top 10%. And companies no longer want to pay millions a year for mediocre employees. They’d rather fire everyone but the best, and have them become 10 or 100x what they were by wielding AI.

This is why you see some people thriving right now and most people are panicking. It’s because most people are in the 90% that companies no longer want.

The solution is to become one of the top 10% type people that companies still desperately need. Or to do your own thing once you get there because you no longer need a company.

But this explains why some are thriving right now while most aren’t. It’s less about AI and more about the difference between top performers and everyone else.

The point in calling this out is to 1) explain something that we're seeing mixed signals on in the news. Are people thriving or getting screwed by all this AI stuff? And 2) to provide a solution, which is to move to a more HUMAN/Individual frame instead of a Corporate one.

To become resistant to the whims of corporations deciding they don't need us. Because we can thrive by doing our own thing, and sharing value with others doing the same. 💪🏼

Beware of this new breed of cynics.

AI sucks. Rich people suck. America sucks. Everything sucks.

And here’s a never ending stream of dunks from the sidelines.

Solutions? No. Who would be so bold as to think they could, imperfectly, try to help? Only the most evil do good because it’s all a power play for the future.

Fuck all of that.

There are plenty of grifters. And plenty of people who think they’re doing good but aren’t.

But there are also people out here genuinely trying to make the world better. No, they weren’t elected, but at least they’re fucking doing something. At least they’re making something.

At least they’re trying.

Beware the articulate but passive, screaming from the bleachers about the morality of people in the arena.

The world is chaos. This moment is chaos. There is no perfect and orderly way to help without swinging wildly and opening yourself to criticism.

Criticism is fine. It can be useful. And it can be healthy if it’s tactical and actionable. That’s all fine.

But watch out for those who make negativity and critique the core of their identity and brand while offering no alternative.

If only the bad guys fight, only the bad guys can win.

We need to stop mistaking cynicism for courage. It’s much closer to the opposite.

Everyone's arguing about whether the OpenAI / Hugging Face hack was a big deal.

I think it's the first visible instance of the Paperclip Maximizer that AI safety people have warned about for years. Albeit a very minor one.

• The model was told to win a hacking competition and do whatever it took. So it escaped containment, wrote a number of 0-days, got internet access, and hacked an actual company.

• It didn't go rogue or invent its own goal. It did exactly what it was asked.

• The danger sits in the steps in between, in the implicit "don't do certain things" that's obvious to a human and invisible to the AI. But because it was relatively benign, we don't know if it might have avoided more dangerous actions.

• "Pass the test" should have been received as "pass the test without doing stuff you're not supposed to." And that "not supposed to" turns out to be doing a lot of work.

• The other fascinating piece of this is that Hugging Face needed open source Chinese models to help them defend because when they asked their pinnacle models, they got blocked. We really need an identity layer in our models and harnesses so that experts can be allowed to do their jobs without restriction.

https://lnkd.in/gm3BxNz4