Emir Ozer

[Emir Ozer @ozermiozer]

dude you should've just moved to the Codex team instead

[gabriel @gabriel1]

3 months ago i resigned from @OpenAI. today we are launching Energy

everyone will soon use AI for all work on their computers, but very few do today even though AI is good enough

our users do days of work in hours. we'll onboard the world to working with ai

[Emir Ozer @ozermiozer]

@gabriel1 dude

[Emir Ozer @ozermiozer]

@theo @dreamwieber try this in a corp device with 5 different security agents hooked into every file operation

[Emir Ozer @ozermiozer]

thanks @sama for making this happen 🫡

[Claude @claudeai]

Beginning July 20, Claude Fable 5 will be included in all Max and Team Premium plans, at 50% of limits.

Pro and Team Standard users will continue to have access to Fable via usage credits, and will receive a one-time $100 credit.

Demand for Fable has been challenging to predict, which is why we rolled it out to subscription plans in stages, extending access several times as we secured additional capacity.

[Emir Ozer @ozermiozer]

damn this place is becoming the new linkedin for tech bros

[Emir Ozer @ozermiozer]

@markletree @coinbase The question is have you seen any degradation in quality when routing to cheaper models or any disruption to your workflows? This feels like definitely where we are heading-sustainable AI usage- and I think we still have a lot of optimization work to do inside harness and context

[Emir Ozer @ozermiozer]

@underalive84 @bryan_johnson You stop liquid few hours before before bedtime

[Emir Ozer @ozermiozer]

@fullbarsmedia @bcherny check your env values for any hardcoded sub agents model , if not I usually prompt to main agent to spawn sub agents with the right model for that task which is a cheap win for reducing cost

[Emir Ozer @ozermiozer]

@fullbarsmedia @bcherny If you are in the ultracode effort mode it loves to spawn an agent army even for a simple task sometimes . the worst combo if the sub agents are also spawned with expensive models like fable or opus which is not ideal and can be overkill for some cases

[Emir Ozer @ozermiozer]

@artemmurzin How are you planning to address the privacy concerns? I would be very skeptical about sharing my whole chat data in clear format with a third party tbh. those chats include sensitive confidential information

[Emir Ozer @ozermiozer]

@xNamuken you missed the point. We don’t want to disable it. The word "workflow" is over-triggering the feature, which is very token intensive. It should be a more deliberate decision

[Emir Ozer @ozermiozer]

@PerceptualPeak @bcherny just fyi - opus 4.8 is the default model in the latest version so you don’t need to specify

[Emir Ozer @ozermiozer]

@youplusai @phuctm97 end to end test doesn’t capture code quality, architectural drift, individual performance, code maintainability etc

How do you validate those aspects without reviewing? Genuine question

[Emir Ozer @ozermiozer]

AbstractHammerFactoryFactory

[Yoonho Lee @yoonholeee]

We just released code for Meta-Harness! https://t.co/OdU7zocdPl

Aside from replicating paper experiments, the repo is designed to help users implement good Meta-Harnesses in completely new domains! Just point your agent at ONBOARDING.md and have a conversation

[Emir Ozer @ozermiozer]

@rsgalloway @nixkristall @unclebobmartin quantum mechanics would argue with that statement

[Emir Ozer @ozermiozer]

@andrewhsandler @FarzaTV graphrag didn’t even exist three years ago

[Emir Ozer @ozermiozer]

In plan mode it jumps into the conclusion and starts building even when the requirements aren’t fully ready. So I usually groom my idea first by brainstorming and designing it at a high level in regular convo. when I feel ready I ask for a concrete plan to get it done in one shot

[Anthony Kroeger @kr0der]

slowly starting to use plan mode a LOT less nowadays

i realised whenever i use plan mode, it generates a gigantic plan and then i dont read it and hit build out of laziness

having a meaningful conversation with the AI agent to discuss implementation feels a lot easier 🤔

[Emir Ozer @ozermiozer]

Codex CLI has gotten really good recently all of a sudden. I wonder what their secret is 🤔

[Frederik Jacques @thenerd_be]

Today is my first day at Anthropic.

Super excited I shipped my first change today, added source maps so debugging is easier.
Can’t wait to show you all what I’ve been working on!

cc: @AnthropicAI

[Emir Ozer @ozermiozer]

@eliadeleo @heygurisingh ai wrote this what are you on about

[Emir Ozer @ozermiozer]

@rezoundous idk about the desktop app but codex cli is terrible