Lillian Ma

event host: black & red only me: open my gmi console to pass the outfit check @yuqih @sheismikoshu

@yuqih Snow boss is checking on my token burn 🔥

@lillian_ma_ I hope Snow was entertained while Claude Code was discombobulating 🐩

pov: me & office dog waiting for my Claude Code to finish solving a million $ question

@yuqih @gmi_cloud @venjia_z I was thinking when did you take the pictures together

Hi, I'm Yuqi! I just launched the ambassador program for @gmi_cloud with @venjia_z! Bit about it: - $200–500/mo in credits to build - funding to host your own meetups + events - your demos featured across GMI - swag Let's connect if you're into building with AI, making content/demos, and running discord or irl events

for my builder friends who are looking for a community, plus free credits every month

@meh_agarwal @nicoleegong @yuqih Nicole got our whole girl group signed up

@nicoleegong @lillian_ma_ @yuqih Tell her to sign up luma.com/cw9b23l4

great benefits when you work at Castro St in MTV so many dessert options during afternoon’s break @yuqih @nicoleegong

@yuqih @nicoleegong I'll meet you at technically single event this Fri 😊

@lillian_ma_ @nicoleegong Cuteeeee we want an intro 😊

@yuqih when will you start posting stories around sf hacker community haha

Hi I’m Yuqi! I recently moved to SF. Bit about me: - live in a growth house - 10k on youtube - used to live in nyc, paris and london - writing a novel on the caltrain - working at an ai startup Let’s connect about agents, ai models, making content and writing fiction

love the brutal comparison

@bj0hn5on @nvidia @gmi_cloud @CrusoeAI @friendliai @baseten @togethercompute @nebiusai @DeepInfra @nscale @LightningAI @Vultr Will we have more events like this? 👀

Inference Co-design Day was a hit! 💚 S/o to all the @nvidia Cloud Partners and Inference Service Providers that came out for a deep dive on optimizing inferencing. @gmi_cloud @CrusoeAI @friendliai @baseten @togethercompute @nebiusai @DeepInfra @nscale @LightningAI @Vultr @SimplismartHQ @digitalocean @Eigen_AI_Labs @inference_net

@gmi_cloud hehe

@lillian_ma_ 👋

Come say hi to me if you’re at our @gmi_cloud Scale Program Demo Day

What's the market availability for Blackwell? Effectively 0% for new buyers. No kidding. I've personally fielded 10x inquiries in the past month around "do you have X GPU," and I feel a lot of founders and builders still haven't touched the truth: there's just no liquid compute capacity on the market right now. 6 months ago if someone told me H100 1-yr contract pricing would surge 40% in half a year (from $1.70 to $2.35/hr) with 36–52 week lead times and capacity sold out through Aug–Sept 2026, I would've thought they were crazy. And it's not even a money game anymore. The liquidity has been locked down by hyperscalers (~$700B capex in 2026), neoclouds and frontier AI labs already. Anthropic is renting from a direct rival's data center because that's where the electrons are. That's the market we're in. So here comes the trend: -For small OSS providers: fewer cloud providers are going to host you. Attracting customized deals from public endpoints will soon become a top-player's game. -TCO matters more. People who actually handle token economics optimization will be valued more on the market. -Industry players are desperately pushing forward an inference-efficiency-driven playbook. -Hard times ahead for any ecosystem built on top of spot instances. -Founders dreaming of Claude lowering their token price should drop the idea immediately. Yes, Opus 4.5/4.6 came in cheaper than 4/4.1 — but Opus 4.7 kept the headline price flat, the new tokenizer can raise effective costs up to 35%, and now they're paying premium rates to a rival for emergency capacity. If you can't justify your unit economics right now, you'll only burn more money per user in the near future. -TTS models will have a huge advantage in this era to justify their margins. Opinions here are on my own.

@yuqih I want a tennis dress then @nicoleegong

@lillian_ma_ GONNA MAKE IT yes I cannot wait

We’ll order some “Gonna Make It” swags with the new logo on

@KranenKyle @nvidia You did a great job today Kyle! Thanks for sharing ❤️

@lillian_ma_ @nvidia Thank you for coming to inference day with us!

Big day 🚀 As early adopters of TRT-LLM, Dynamo, and NIM, we’re at @nvidia’s Inference Codesign Day meeting the team IRL. As an inference infra provider, partnering with NVIDIA to bring world-class inference to our customers is exactly where we want to be. Heard TRT-LLM is expanding coverage for visual generative models 👀 the road ahead is going to be wild. @gmi_cloud @bj0hn5on @ReneeYao1 @NVIDIAAI @NVIDIAAIDev

@thejessezhang @cognition @cursor_ai @perplexity_ai @AnthropicAI @thinkymachines @midjourney @tryramp @xai @EvidenceOpen @DecagonAI I want to sign up but my boss said nono don’t lose our company face

We're hosting a sick poker tourney with 24 top players from @cognition, @cursor_ai, @perplexity_ai, @anthropicAI, @thinkymachines, @midjourney, @tryramp, @xAI, @evidenceopen, @DecagonAI, etc 1 player repping each company. Prizes, live sushi chef, limited viewing spots, RSVP 👇

@radixark @Accel @sparkcapital Congrats!

Today, we are thrilled to officially launch RadixArk with $100M in Seed funding at a $400M valuation. The round was led by @Accel and co-led by @sparkcapital. RadixArk exists to make frontier AI infrastructure open and accessible to everyone. Today, the systems behind the most capable AI models are concentrated in a small number of companies. As a result, most AI teams are forced to rebuild training and inference stacks from scratch, duplicating the same infrastructure work instead of focusing on new models, products, and ideas. RadixArk was founded to change that. We are building an AI platform that makes it easier for teams to train and serve the best models at scale. RadixArk comes from the open-source community. We started with SGLang, where many of us are core developers and maintainers, and expanded our work to Miles for large-scale RL and post-training. We will continue contributing to both projects and working with the community to make them the strongest open-source infrastructure foundations for frontier AI. We would like to thank our long-term partners, contributors, and the broader SGLang community for believing in this mission. We're also grateful to @Accel and @sparkcapital, NVentures (Venture capital arm of @nvidia), Salience Capital, A&E Investment, @HOFCapital, @walden_catalyst, @AMD, LDVP, WTT Fubon Family, @MediaTek, Vocal Ventures, @Sky9Capital and our angel investors @ibab, @LipBuTan1, Hock Tan, @johnschulman2, @soumithchintala, @lilianweng, @oliveur, @Thom_Wolf, @LiamFedus, @robertnishihara, @ericzelikman, @OfficialLoganK, and @multiply_matrix among others. Thanks for the exclusive interview with @MeghanBobrowsky at @WSJ about our vision.