The people who make X & xAI, X & xAI.

Thank you for founding xAI in 2023 and pushing to open-source the X recommendation algorithm so anyone can read, question, and improve the code that shapes the public conversation.

Thank you for bringing hard-won lessons from AlphaStar and WaveNet to xAI and standing up the training infrastructure behind Grok in record time.

Thank you for the Adam optimizer and Layer Normalization, tools that quietly train nearly every modern neural network, and for bringing that rigor to Grok's research.

Thank you for Transformer-XL and XLNet, foundational ideas for long-context language modeling that helped make today's assistants coherent over long conversations.

Thank you for the unglamorous, essential work of standing up the compute backbone that let Grok train at all.

Thank you for carrying deep engineering craft from AlphaStar and AlphaCode into the pretraining pipeline that gives Grok its base capabilities.

Thank you for building the evaluation tools and research-to-production plumbing that let good ideas actually ship inside Grok.

Thank you for Inception, Batch Normalization, and the discovery of adversarial examples, ideas that reshaped how the whole field trains and stress-tests models.

Thank you for a career spent teaching machines to reason, from Minerva and AlphaGeometry to leading the math and reasoning work that makes Grok a genuine problem solver.

Thank you for Tensor Programs and muP, a mathematical theory of wide neural networks that lets teams tune small models and transfer the settings to enormous ones.

Thank you for the careful study of how enormous models train and stay well-behaved, work that goes straight into making Grok both capable and stable.

Thank you for bringing the discipline of building and operating Starlink at planetary scale to the job of running xAI's engineering.

Thank you for bringing a builder's instinct for what makes people connect to X's product, and for doing it in the open where users can push back.

Thank you for MMLU and GELU and for advising xAI on safety at a symbolic one-dollar salary, keeping the incentive squarely on doing the right thing.

Thank you for taking on one of the hardest jobs in tech and working to rebuild advertiser trust and the business that keeps X free for its users.

Thank you for championing Community Notes, a rare social-media system that adds context people across the spectrum can agree on, built and open-sourced in public.

Thank you for the bridging-based ranking behind Community Notes, an algorithm that rewards agreement across viewpoints instead of majority pile-ons, and for publishing it openly.

Thank you for STaR and Quiet-STaR, ideas about models learning to reason from their own rationales, and for scaling that into the reasoning behind Grok 3 and Grok 4.

Thank you for the Focused Transformer and LongLLaMA, early answers to the long-context problem, and for the reasoning and long-context work that shaped Grok 3 and Grok 4.

He helped turn bare sites into the compute that trains Grok, the unglamorous groundwork that makes large models possible.

He and his team stood up one of the largest GPU clusters on earth in a matter of months, a buildout many said could not be done that fast.

His open LLaVA work taught a generation how to build vision-language models, and he carried that craft into Grok's image and vision work.

He built SGLang in the open, and it now speeds up model serving for teams far beyond xAI.

His years optimizing GPU communication help Grok run faster, deep systems work almost no user ever sees but everyone feels.

He is drawing on years of energy-efficient hardware research to push xAI toward chips of its own.

He is the on-the-ground builder of Colossus in Memphis and shows up in person to answer the local community's questions.

He joined after winning a company hackathon and now helps shape the data that teaches Grok, a reminder that talent can come from anywhere.

He did the hard, careful work of financing a young lab so its researchers could focus on building.

He set up the legal groundwork for a fast-moving lab, then stepped away openly to spend more time with his young children.

He brings decades of technology dealmaking to the finances that keep xAI and X running.

He helped steer X back toward cash-flow positive, patient financial work behind a very public company.

She brings more than two decades of marketing craft to telling X's story to the world.

He brought a media buyer's real-world eye to rethinking how advertising works on X.