No one wants to hear this but these guys are making the next BRK type fandom

I haven’t seen AI as a bubble in some time, but I really question those with a strong opinion that it isn’t if you are not up close and personal with how knowledge workers are using these tools.

Another thing: token based measures of demand, usage, etc are worth zero. Nothing. Absolute hogwash.

I’m watching an agent fail at a 50 lines of python task and using >100k tokens to do it. Oh, it also takes 10 minutes now for some reason.

I feel like it’s under-reported how variable AI output is becoming.

Would Google have become as dominant if every 3 weeks, searching for a restaurant yielded extremely different quality results?

Who asked for the talks to be on Saturday morning?

Don’t think this ceasefire lasts, either

Idk what’s gonna happen but I was raised right: to buy calls when the president threatens annihilation

Once you get “Trump reciting the Shahada” into your mind, it never goes away

I love how stupid we’re all getting

a cartoon character with a nick logo on the bottom left ALT: a cartoon character with a nick logo on the bottom left

*Vibes based analysis*

This seems like an environment where an AI shoe could drop and really get people in trouble.

Once you get “Trump reciting the Shahada” into your mind, it never goes away

It’s so sick that you can tweet false and misleading information just because you have a fake job like “Secretary of Energy”

Comment 1999 below to get my report on how to survive the next 6 months

I don’t know too many Waymo skeptics who have stayed that way after their first ride, but we have gone from a generation of tech whose consumer surplus is so immediate and obvious that people have forgotten what it’s like to market tech without it.

Think this is roughly right, but the OS models also get some low end SMB-like use cases.

Totally agree, software underperformance is a big deal

2026 Predictions:

* Consumer preferences will go from hostile to only mildly negative for AI in media/outputs.

* The Year of AI Psychosis

* We’ll keep ascending AI benchmarks, keep hyperbolizing, and keep wondering why it doesn’t quite get the job done.

Perhaps thats not useless.

Think it’s implicit in many views that sees LLMs doing things where certainty of execution and execution profile is super important.

My keys are valuable because I know the lock will 100% open when I turn it. If not, I know how to fix it. If I introduce an LLM to my lock, what do I gain?

A few observations on AI that have been interesting to me:
* A significant rise in engineers calling the recent models AGI.

* A general investor reluctance to admit there’s a bubble, even though most of the industry admit one

* The implication that there’s no value in deterministic processes.

Every now and then I see this chart and think “huh, they just kinda put it all out there on a quarterly basis, don’t they?”

“We’ll see how [obviously bad thing] turns out in the data” is such a good online position to always hold because it lets you be both incredibly pompous and so late to everything that you can pretend “but that’s obvious” when you turn out to be wrong.

b438f3ce24f2aa1318128c47fc7a5b61

Do not say this in polite company

A brain fully imprisoned by the Laffer curve and handcuffed by tax policy.

Reversing tariffs will be the *easiest* of the tasks ahead for the next president.

Lindsey Graham wining about Norges selling CAT

Selling an equity stake in an American business is a punishment, and not something that a foreign investor can do without consequence.

Thank god for this pro-business environment.

“It’s Joe’s conspiracy to rehabilitate this woman’s legacy and sanewash overreaching political actors to suit his agenda”

Joe:

The serious person take on all this is that somehow SCOTUS and other institutions thought the Fed would somehow still be off limits? There was an implicit bet or bargain?

C’mon. You’ll be marks until your last day.

What makes you say that? Def seems to be doing fine but thought their trade down effects + unique tariff situation might make them unique for 2H.

A few things in the org structure here stick out as strange to me. It’s beginning to seem like Zuck promised each of these major hires their own fiefdom, and they’re more likely to go to war than accomplish their goals.