Sean Blahovici

[Sean Blahovici @blahovici]

@gravityfast_ @capsraunak Using ai requires no true skill though.

[Sean Blahovici @blahovici]

@vineerpasam Nah.. He isn't actually learning. You learn by doing. Writing code. Struggling to solve a problem. Trying and failing over and over until you succeed. Reading an LLM's response might teach you something in the moment, but you won't retain it, or be able to use it.

[Sean Blahovici @blahovici]

@MatthewBerman @bot I definitely would be interested in trying Grok bot!

[Sean Blahovici @blahovici]

@TTrimoreau 0%. In fact negative, because I can build without it. Without any AI. I would prefer it.

[Sean Blahovici @blahovici]

@anthonygore Block out 60 to 120 minutes per day of AI-free, cognitively demanding work. Brings back flow and satisfaction, and reduces the guilt of having an AI do most of your thinking for you.

[Sean Blahovici @blahovici]

@ClaudeDevs Auto mode is great inside a sandbox, paired with pre tool use hooks that block reading and writing outside of allowed directories. May not be bulletproof, but I feel relatively at ease leaving Claude unattended.

[Sean Blahovici @blahovici]

@unclebobmartin So you do all this without reviewing the final output, but you watch the agent work. Does that mean that you run one of them at a time and watch the output so you get a general feel of the direction it's going, and that gives you confidence to skip final review if tests pass?

[Sean Blahovici @blahovici]

@beeeeeeau @rand_longevity Yeah but scaling production of these types of robots is very difficult as it requires huge factories, rare earth metals, etc. You need a lot of robots to replace blue collar workers entirely.

[Sean Blahovici @blahovici]

@unclebobmartin @AdamRackis I consider you an amazing teacher and learned much from you. But I am disappointed by the idea that you do not read the code agents produce. Do you read the tests that are generated at least?

Perhaps I need to check out your clean AI series..

[Sean Blahovici @blahovici]

@iwasrobbed @NickADobos The primary value of software is its ability to evolve over time without an ever increasing time to stability. The secondary value is that it works right now.

Well designed systems matter. Code matters.

[Sean Blahovici @blahovici]

@rishivenkat30 @claudeai @AnthropicAI Same

[Sean Blahovici @blahovici]

@MrBeast 3 523 789

[Sean Blahovici @blahovici]

@claudeai Claude Code remote control has been in research preview for months. Can you please graduate it to stable so we can use it within our corporate team? Thanks!

[Sean Blahovici @blahovici]

@unclebobmartin Do you find that agent-assisted development is eroding your ability to write code manually and your capacity for deep work, seeing as you must wait for the agent to finish it's work and are incentivised to context-switch often?

[Sean Blahovici @blahovici]

@unclebobmartin Nice! Can we get some availability in Amazon Canada plz

[Sean Blahovici @blahovici]

@sama Deepseek r1 provides reasoning with search, will o3-mini also support search & reasoning?

[Sean Blahovici @blahovici]

@sama Search with structured outputs in the api please!

[Sean Blahovici @blahovici]

@OpenAI How come my ChatGPT team account members cannot access Sora? The team subscription is more expensive than a personal subscription...

[Sean Blahovici @blahovici]

@EricLDaugh How about we work with them instead of offering counter threats?

[Sean Blahovici @blahovici]

@sama I have two different chatgpt plus accounts and Sora says I need to upgrade. I live in Canada