Steve

not a bot

[Steve @SteveWesterhout]

@GlowSurfing Toronto concert after Sars... ACDC blew the doors off the Stones

[Steve @SteveWesterhout]

@DanielMiessler really not a surprise that pro gun people use the same logic for AI.

[Steve @SteveWesterhout]

@Arianikle

[Steve @SteveWesterhout]

@sirius_devops @DanielMiessler 🍍

[Steve @SteveWesterhout]

@mike_clemmons @DanielMiessler you think the US was the only one to help?

[Steve @SteveWesterhout]

@DanielMiessler mostly still on opus 4.6, 4.7 wasn't worth the tokens, don't have 4.8 yet. I'm also doing very well with my setup based off of PAI.

[Steve @SteveWesterhout]

@BrettKollmann @BruceExclusive dumb ones in every country tend to have the loudest voice.... that's not an American problem... although dumbest and loudest may be your fearless leader which does not help

[Steve @SteveWesterhout]

@AlbertBreer 2nd that on Canada's behalf. It may not be our #1 sport, but football (even the Canadian version) is very different from soccer

[Steve @SteveWesterhout]

@kmonii1776 @Polymarket and the us military will spend 10x on the missile to shoot down the helicopter vs the cost of the helicopter ✅

[Steve @SteveWesterhout]

@DanielMiessler been using PAI daily now for months. everyone always asks me why I get better AI results then them

[Steve @SteveWesterhout]

@DanielMiessler well said

[ᴅᴀɴɪᴇʟ ᴍɪᴇssʟᴇʀ 🛡️ @DanielMiessler]

I'm really confused about all this hate towards Anthropic about sending (supposedly) mixed messages.

If you were an AI company and your latest model popped out and was WAY better than you anticipated on cyber, your reaction would likely be:

1. Freak out a bit

2. Realize how much it could help attackers in the wrong hands

3. Realize this is just normal model progress for ALL vendors, so time is limited

4. Realize it could also help defenders get ready for the inevitable

5. Realize it would be great marketing to talk about how powerful it is

6. Come up with a plan to give it to defenders first, to clean up a bit before every provider has something as good

7. Release a safer version of the more powerful one, but with more safeguards, while keeping the more dangerous one still restricted

This is PRECISELY what they did, and it seems super damn logical and what most ethical companies would do in their place.

It's the same with them pushing forward with AI progress while also warning about the potential downsides?

What would you do? Stop? Shut down the company? Stop developing the models and the harnesses and applications and hope that your competitors and China do the same?

No. Most ethical companies would NOT do that.

What they would do is:

1. Continue innovation

2. while still releasing tons of safety content and going on podcasts and doing conferences to talk about the messaging

3. while SIMULTANEOUSLY doing marketing for the actual company because who doesn't do marketing for cool stuff the company is doing?

If you look at the whole of it, I don't see a single bit of unethical behavior here.

Of course they're doing marketing of the capabilities. That's marketing's job.

And of course they're still innovating; stopping doesn't help anyone unless everyone stops at the same time.

If you think they did something wrong, imagine yourself in charge of all of this and tell me what you'd do differently. The vitriol makes no sense.

[ᴅᴀɴɪᴇʟ ᴍɪᴇssʟᴇʀ 🛡️ @DanielMiessler]

Another major AI vibe shift is happening.

The tech is moving so fast that our collective reactions are emotionally exaggerated.

In mid 2022 most didn't think anything could pass the Turing test.

We get chatbots in 2023.

We get agents in 2024. But nobody trusted AI for coding.

We get Claude Code in 2025, and in one calendar year we go from "I'd never let AI code for me", to "I let it write the code but I review it", to "I don't even write code anymore."

In late 2025 and early 2026 most boards and CEOs start demanding everyone use AI. For what they're not quite sure. The vibe is "we need to immediately replace workers with AI." People think everything will be AI inside companies within 18-36 months.

And now in mid 2026 the pendulum is coming back. "Oh crap that's hard and expensive. We better slow down and rethink this." People start thinking this was a giant miss, and that human workers will never go away.

The pendulum swing from "it can't code" to "I don't code anymore" happened in less than a year.

And now this one has gone from "AI will replace all human workers" to "AI can't replace human workers" in less than 6 months.

Pardon the cringe analogy but I'm reminded of when someone is dying of cancer (in this case I think the thing dying would be the old way of work being done).

When someone gets a major / terminal cancer diagnosis there is tons of uncertainty. You keep testing and testing, and you often have two groups of people.

We've already been told it's terminal, but whenever there are positive responses to a treatment, or positive test results, they conclude that "It looks like we beat this thing, you'll probably live another couple of decades."

And there's another group who, upon hearing any negative result or analysis, says they could pass any day, or any week. They'll be gone in less than a few months.

I've seen this structure play out multiple times.

1. It'll happen in less than 2 months.
2. It'll never happen.

And sometimes people in one group might switch to being in the other group.

But what I've seen basically every time is that it ends up being something in the middle. It actually doesn't happen instantly like the fast group said. And then right when it starts to look like things are going to be fine, it happens.

Somewhere in the middle.

I hate this analogy for obvious reasons, but I feel like we are collectively stuck in this mode when it comes to the inevitability of change from AI.

The combination of "really bad" with "don't know when" causes a special kind of anxiety and reaction in us humans.

Also, this model doesn't necessarily tell us what's happening in AI because it depends on what you use as the substitute.

My intuition is that the thing at risk in our current situation is "most work being done by humans". That is the thing that now has limited time.

And you can find all over Twitter the people who think 20% unemployment or whatever will happen in 1-2 years. And an equal number of people who think this is a psychosis that will soon end. Or that it'll take 10 or 20 years instead.

This reminds me a little of regression to the mean. It's something like find the median number for all the guesses.

But that method wouldn't have been correct in early 2022, so who knows.

I just think this RTM mental model might be useful in thinking about what could happen and how long it might take.

[Steve @SteveWesterhout]

@NateGearySports or none of the above because if you're are not a US citizen you risk getting thrown in an ICE concentration camp 🤷

[Steve @SteveWesterhout]

@joemillerwired it's the red hair @joemillerwired ..... keeps you young

[Steve @SteveWesterhout]

@DanielMiessler I can't find anything there i significantly disagree with. fully agree that being a builder/creator and being good with AI is the best path forward. I worry for those that are not creators.

[ᴅᴀɴɪᴇʟ ᴍɪᴇssʟᴇʀ 🛡️ @DanielMiessler]

A few things here.

First, I like @atmoio a lot. I love the style of thinking out loud. Please keep it up dude.

As for this argument, I have a few problems with it.

First, I think this is (unfortunately) at least partially a direct response to audience capture. Mo got popular as an anti-AI guy, and then he does his previous video, and he got flamed.

So I worry a little bit that some of this is him trying to make things fit. I respect his intellectual honesty so I'm not SUPER worried, but...rent is too damn high. Everyone is vulnerable to audience capture so I don't fault him if it got him a bit.

Ok, now the ACTUAL argument. And this isn't just a Mo thing.

I think we should use definitions of things that matter to humans. And that are not just useful to humans, but use-able. Usable. Usable definitions.

The definition given here for intelligence is highly esoteric. Basically it's the shape of the universe or something. Beautiful. I can see it. But hardly anyone can or will use that definition.

Let's back up and think about why we even care about any of this.

- Why do we care about AI?
- Why are people worried?
- Why were his videos so popular when he was bashing AI?

It's because many see AI as a massive threat to people.

Why is it a threat? Because it can do some kinds of work that humans used to do. And replacing human jobs is serious shit.

That's why we care.

So the topic of conversation, I would argue, should not fucking be intelligence being the shape of the universe or some shit. I get it. That's cool shit. Love it.

But it's not what people care about, and therefore what matters.

What matters is how well AI can do things that humans do, and get paid to do, and find meaning in doing. Full stop.

So I think a better definition is something like the ability to take a problem and a goal and get a desired outcome. Dead simple. Practical.

Let's use something like that. Please.

Now for the second piece of his argument.

He talks about how you can't cheat steps. Evolution took billions of years, so therefore this thing we just made can't be intelligence.

This is super easy to prove wrong.

It took evolution millions or billions of years to make eyes too, but we can build optical sensors in a matter of minutes.

Even better, if we find a new sensor that some animal has that we didn't know about, like being able to smell some disease, we can reverse engineer that thing and potentially make one.

And once we know how, we can then make that sensor in minutes.

He's saying we can't cheat, but that's literally what education and culture and technology is. It's doing what takes billions of years in minutes or days or a couple decades.

So:

1. Mo is awesome and I hope he keeps going
2. I hope he resists the audience capture, because he's too good a mind to get wasted on forcefully arguing only one position
3. Let's use a practical definition of intelligence that matters to humans, and that's usable to humans
4. The idea that we can make our own intelligence technology that mimics what took evolution billions of years is literally a thing we do all the time with technology

@atmoio Appreciate you. Keep pushing.

[Steve @SteveWesterhout]

@BruceExclusive well said

[Steve @SteveWesterhout]

@FoxNews or send those adults to Antarctica to dine, I doubt there are kids there.

[Steve @SteveWesterhout]

@github gitlab
SVN
CVS
PVCS
something on the mainframe that was so long ago I can't remember....
before that a hope and a prayer

but that's over 29 years 😁