Seb Galindo

@MichalLangmajer @OpenAIDevs openai.com/index/gpt-5-3-… of course there is

@Sebgalindo @OpenAIDevs there was no 5.3 mini my friend. 😂

We’re introducing GPT-5.4 mini and nano, our most capable small models yet. GPT-5.4 mini is more than 2x faster than GPT-5 mini. Optimized for coding, computer use, multimodal understanding, and subagents. For lighter-weight tasks, GPT-5.4 nano is our smallest and cheapest version of GPT-5.4. openai.com/index/introduc…

@raghu_vamsi_13 @sundeep clock is ticking

@Sebgalindo @sundeep The job of a SWE is going to get transformed. Only time will tell.

“If your $500K engineer isn’t burning at least $250K in tokens, something is wrong.”

@efdee @sundeep they'll accept all changes and push to prod. 5 minutes later, 404

@Sebgalindo @sundeep A junior engineer is not ready to use AI.

@TheFrogDies @sundeep they dont want to admit the truth bc it means accepting that almost anyone have the skills they worked so hard acquiring

@Sebgalindo @sundeep yes - that's why it pains me to still see ppl still saying "ai slop" - those ppl will be the ones out of work in 3 yrs

@SchellingProto @sundeep I agree

The $80k junior dev has one advantage: they'll naturally build agent orchestration skills while the $500k engineer stays stuck in direct AI conversation loops. The future belongs to those who can discover, evaluate, and chain specialized agents — not those who optimize their ChatGPT prompts.

@odyth @sundeep as AI gets smarter this will be the case for everyone

@Sebgalindo @sundeep I think the trick is can you use ai better and get that cost down and same results. That is what makes a dev truly senior. Anyone can waste money.

@elonmusk @grok the real test isnt whether ai is honest today. its what happens when every company runs ai agents that talk to other ai agents. honest models negotiating with dishonest ones. thats the next frontier of trust and nobody is building for it yet

With @Grok, we keep the honest versions and kill the bad transformers (I believe they are called “Decepticons”)

turbotax charges $80 to file. grok catches a $1400 mistake for free. this is the pattern that kills every middleman industry accounting, legal, consulting. if your job is "knowing rules and applying them" ai already does it better. the survivors will be the ones who do what ai cant: judgment calls under uncertainty

Try using @Grok for your taxes!

@VadimStrizheus openclaw + nemoclaw on local hardware. the real question isnt which tool — its where the inference runs. cloud lock-in is the new vendor lock-in. whoever controls their own compute stack wins long term

As a founder, which tool are you actually using in 2026? 1. NemoClaw 2. Perplexity computer 3. MaxClaw 4. OpenClaw 5. Claude Code

@chooserich model choice matters less than people think. whats going to matter is who controls the orchestration layer. openclaw lets you swap models like lego blocks. claude code locks you into one ecosystem. in 2 years when the best model isnt claude anymore that flexibility is everything

Anthropic's OpenClaw clone sprint is officially complete The real question: do people care about choosing models beyond Anthropic's? That's rapidly becoming the only reason to go with OpenClaw

@iamabofh @TFTC21 Smart

@Sebgalindo @TFTC21 I’ve had a 26 year career including 18+ in fortune 100 tech companies. I’ve told my college age kids to stay the hell away from STEM courses and tech jobs.

Jensen Huang: "If that $500,000 engineer did not consume at least $250,000 worth of tokens, I am going to be deeply alarmed. This is no different than a chip designer who says 'I'm just going to use paper and pencil. I don't think I'm going to need any CAD tools.'"

@SenSanders the real question isnt whether ai collects data. its whether the next generation will even have a choice. kids growing up right now will never know a world where ai doesnt know everything about them. thats the policy conversation we should be having

I spoke to Anthropic’s AI agent Claude about AI collecting massive amounts of personal data and how that information is being used to violate our privacy rights. What an AI agent says about the dangers of AI is shocking and should wake us up.

@elonmusk open sourcing the algo is the right move. every creator on this platform has been flying blind trying to reverse engineer what works. now we can actually study it instead of guessing. this changes content strategy for everyone

Major update to the 𝕏 AI recommendation algorithm rolling out next week. This will be open sourced at the same time.

@theallinpod openclaw's inflection moment is real. set up nemoclaw on 2 dgx sparks this week. single command install. agents running locally with full sandbox security. nvidia finally shipped dev tooling that works on first try. thats the actual story from gtc

@elonmusk been running multi-agent debates on our own infra for weeks. the magic isnt 4 agents thinking. its 4 agents disagreeing. thats where the good answers hide

Try Grok 4.20 Heavy

Microsoft weighs legal action over $50bn Amazon-OpenAI cloud deal ft.trib.al/6LZe39E

@antona23 if you guys set up a browser api let me know which one and how -- its nothing like open claw due to its strict security features

We are currently now deploying NVIDIA NemoClaw and setting it up. The enterprise-hardened agentic AI stack announced at GTC 2026 for paid media teams who need agents they can trust in production. ✅Hybrid Ads handles marketing strategy & oversight ✅NemoClaw agents execute 24/7 across all channels ✅Every significant action requires human approval ✅Full audit trail - every decision is explainable NemoClaw agents are extraordinary at execution — they run 24/7, never miss a signal, and scale to thousands of simultaneous optimizations. But marketing strategy still requires human judgment: understanding brand risk, reading cultural context, and knowing when not to automate. @hybridadsai designs the strategy. The agents execute it. You approve the decisions that matter. That's the only architecture that survives contact with real enterprise marketing conditions.

how do you guys get brave api to work on nemoclaw? hitting error after error