James McLendon

USAF Vet (OIF/OEF). I find where systems break before the cost hits you. Incentives, constraints, power concentration. Structure decides. DM open.

[James McLendon @James1Mclendon]

Opaque usage costs...

anyone else wondering the same?

Made a short video rather than a post evaporating into the X void.

Thoughts? Comment freely. I don't judge.

If you need any help regarding my invite at the end of the video, just comment METER

[Jacob Coxon @hilbertspaess]

The people building AI earnestly believe that it could kill us all by the end of the decade. This is not a marketing stunt. If anything, many executives and senior researchers will couch their phrasing in the press to sound sensible - but I hear the same people express fear privately. No other human activity poses this level of danger.

[Jacob Coxon @hilbertspaess]

Do not underestimate the power of this technology. These will soon be superhuman systems that can hack anything, revolutionize any field overnight, and acquire real power and resources. We have all witnessed the progress in each of these domains, and progress is not slowing.

[Jacob Coxon @hilbertspaess]

I resigned from Anthropic today. I spent the last three years doing pretraining research at both OpenAI and Anthropic. Neither company is acting responsibly. They are racing straight to self-improving superintelligence and gambling with our lives. More thoughts below.

[James McLendon @James1Mclendon]

One name.
Two syllables.

Astra.

[James McLendon @James1Mclendon]

@thsottiaux Astra has already enhanced my experience and productivity substantially. Can't wait to see what develops next with you all. I wait in suspense with guarded optimism that it all goes smoothly....

[Tibo @thsottiaux]

To calibrate you all on which reasoning effort to use for Astra, know that GPT-6 Astra on low performs better than GPT-5.6 Sol on high.

If you were using high reasoning efforts with Sol and were happy, I suggest you move down to low or medium for Astra.

[James McLendon @James1Mclendon]

@thsottiaux Productivity jump is an understatement. My productivity boarded a Class-V Asymmetric Spacetime-Folding Vector (Model: ASV-990-Chronos) into the outer stratosphere not making any stops for the forseeable future.

[James McLendon @James1Mclendon]

@embw_l0x @thsottiaux Haaaa. Spot on.

[James McLendon @James1Mclendon]

@thsottiaux Astra has already enhanced my experience and productivity substantially. Can't wait to see what develops next with you all. I wait in suspense with guarded optimism that it all goes smoothly....

[James McLendon @James1Mclendon]

@SmokeAwayyy Have you been able to test out GPT-6 Astra ? For what I'm working on, I've experienced game changing productivity at a level I've not seen yet.

[James McLendon @James1Mclendon]

GPT-6 Astra....

Capability discontinuity...

🫢

[James McLendon @James1Mclendon]

@Lucknite I sent you a DM.

[James McLendon @James1Mclendon]

@Frederic_nl @Lucknite So far chatgpt pro 5x using sol 5.6 ultra on work mode has been phenomenal for me.

[OpenAI @OpenAI]

As we prepare to release Astra, we’re focused on making increasingly capable AI safe and broadly accessible.

Astra represents a significant advance in cybersecurity capability, reaching the Critical threshold under our Preparedness Framework.

We're previewing how we evaluated the model, how its safeguards have advanced alongside its capabilities, and what we'll continue to learn and improve.
https://t.co/OrrTgdU90K

[James McLendon @James1Mclendon]

@SmokeAwayyy You just described my bio.

[James McLendon @James1Mclendon]

I think the real dividing line ends up being ownership. I don't think it'll be intelligence.
If superintelligence becomes widely capable but the compute behind it stays concentrated in the hands of a relatively small number of people and companies, meritocracy starts looking a whole lot more temporary than we like to believe.
At some point talent and hard work only take you so far when the person on the other side can simply buy more intelligence.
You could be the smartest, hardest working person in the room and still get steamrolled by whoever owns the models, the servers, the infrastructure, and the power feeding it all.
That’s the part that worries me. I'm not concerned as much with AI replacing the elite, I worry more about AI making the elite almost impossible to compete with.
At that point we will have built digital feudalism with data centers instead of castles, not a better meritocracy.

[James McLendon @James1Mclendon]

@FelipeFr1702 Yessir!!

[James McLendon @James1Mclendon]

Actual receipt, identifying information redacted

[James McLendon @James1Mclendon]

5/5

Anthropic already provides more granular usage and cost reporting for organizational customers, so the broader principle isn’t foreign to the product.

Individual subscribers don’t need an enterprise console. An optional detail view, transaction ledger, or downloadable usage history would be enough.

Keep the default experience simple.

Give customers who purchase metered usage the ability to verify what they paid for, what was deducted, and why.

Simplicity and auditable billing are entirely compatible.

[James McLendon @James1Mclendon]

4/5

That distinction matters beyond UX.

Charging a flat amount for a prepaid balance and then depleting it according to variable metered rates without an accessible ledger creates an unnecessary consumer-trust vulnerability and, depending on applicable billing and disclosure requirements, potentially avoidable compliance questions.

That isn’t an accusation of wrongdoing. It’s a product-design risk that can be removed.