steveweed.base.eth

Founder @circuit_tv and Generative AI engineer, blockchain enthusiast, live stream producer at Builder on @Base.

[steveweed.base.eth @steveweed_]

I was fortunate enough to be featured as a guest on the @AIPlaybookRavi podcast!

We talk about my career, process to building products, and how I approach generative AI projects like @circuit_tv

https://t.co/fLnL5vzVLM

[steveweed.base.eth @steveweed_]

Timing your weekly Codex reset down to the minute without changing your work habits is the ultimate flex.

[steveweed.base.eth @steveweed_]

@designwkarthick I was under the impression it was the opposite...plan with powerful models, execute with slightly cheaper ones, review with powerful ones. I've posted about this - am I wrong here?

[steveweed.base.eth @steveweed_]

Is anyone else getting the dreaded, 'Sorry, I should have caught that sooner.' replies from @OpenAI GPT-6 Astra?

The last two days it feels nerfed, and has been responding like this and unable to give quality guidance or recommendations.

[steveweed.base.eth @steveweed_]

@pfernan95dev I have been for months. Since around May.

[steveweed.base.eth @steveweed_]

@solopribuilds Gemini? 😂 maybe someday

[steveweed.base.eth @steveweed_]

@NickpxJ Effort control. Astra 6 high for review, 5.6 Sol Max for coding

[steveweed.base.eth @steveweed_]

@moonfarm_dev I'm on the 20x...so, what happens to me? Do I get downgraded?

[steveweed.base.eth @steveweed_]

I have settled in with @OpenAI GPT 5.6 Sol Max as my worker, and GPT-6 Astra High/Extra High as my reviewer.

This has been a solid combo, keeps limits in check and allows for good use of each models strengths, in my opinion and workflow experience.

[steveweed.base.eth @steveweed_]

@Manixh02 Done, done and done at https://t.co/bjeGqJ68oO

[steveweed.base.eth @steveweed_]

@vineerpasam ALL the time.

[steveweed.base.eth @steveweed_]

@0xPaulius What's wrong with it and what would you recommend instead?

[steveweed.base.eth @steveweed_]

I've come to like using GPT-6 Astra for audit tasks. It does really well with those, and reviewer tasks.

Still not a fan of writing code with it, and still use 5.6 Sol Max as my coding model, with Astra as my copilot.

[steveweed.base.eth @steveweed_]

@oldstackjournal That's me!

[Wana @Wanagg_]

Wana is live.

A prediction layer for CS2 streams. A question appears mid-round — "Will they plant?" — you call it before it lands. No money, no odds. Just your streak, your accuracy, your rank.

Closed pilot on Twitch now.
https://t.co/TVByopUCp1

[steveweed.base.eth @steveweed_]

@Ananth7e Not for me, damage is still done and I literally upgraded during all this, unbeknownst to me

[steveweed.base.eth @steveweed_]

@AIandDesign @thsottiaux Me too man almost instantly after an upgrade, smh.

[steveweed.base.eth @steveweed_]

@BradGroux @OpenAI @OpenAIDevs I literally did the same thing; I JUST upgraded to $200/mo from 5x and this happened in the first 30 minutes. @thsottiaux @victornunez

[steveweed.base.eth @steveweed_]

@nickbaumann_ My usage seemed to drain heavily after the update this morning; upgraded router 20x and lost 24% almost instantly

[steveweed.base.eth @steveweed_]

@thsottiaux I just upgraded from 5x to 20x Pro, and 24% of my usage vanished almost instantly.

Before upgrading, this morning, I had 21% and it lasted me a good few hours. This was about 30 mins of light work.

Please help!

[steveweed.base.eth @steveweed_]

@victornunez I just upgraded from Pro 5x to 20x, worked on some light tasks for an hour, and used 24% of my weekly. Not even all Astra either. This is not right!!