Your curated collection of saved posts and media
🚨BREAKING: Jensen Huang just DECLINED Sen. Warren's invite to testify Thursday >Sen. Warren: You must testify under oath about NVIDIA's China sales >Jensen Huang: NO https://t.co/Eg7SX4kkpH
"Korean founders move fast, build with conviction, and don't wait for permission." 🫰 The energy at AWS Seoul Summit was incredible. Law&Company's SuperLawyer, powered by Bedrock, achieved 97.2% efficiency and won the Global LegalTech Awards. They are the first Korean company to do so.
Made in Seoul with AWS: Jaesung Jung, Co-founder & Deputy CEO, Law&Company. By building and running core services on AWS, this legal tech startup is helping lawyers work efficiently and productively. Competition is intense in Seoul! Jung discusses the benefits of Amazon Bedrock and the opportunities and challenges of operating in such a fast-moving city.
Seeing is no longer believing. Reality Defender (@DetectDeepfakes) is stepping in. Reality Defender is leading a cultural shift, helping businesses detect deepfakes in real time across voice and video. Built on AWS, the startup is scaling trust and security for financial institutions, governments & global organizations navigating the AI era.
We cooked. We ate. Literally (and figuratively). @Techweek_ in Boston had us sharing insights, ideas, and culinary experiences. At this exclusive Chef's Table event, startup founders and leaders were joined by award-winning celebrity chef Karen Akunowicz for an evening of connections, conversations and hands-on pasta-making. Sherry Karamdashti, GM of North America Startups at AWS, shares why moments like this matter and what founders are really thinking about right now.
Made in Seoul with AWS: Minyong Lee, CEO & Founder, Grey Box. In Seoul, “the pace is high and everything is local,” helping this startup connect with customers and innovate at speed. Minyong Lee shares his experience of building on AWS and growing a business in such a dynamic city.
It's never been faster for a startup to go from idea to revenue. Today, we're building on that momentum. Introducing AWS Startup Advisor + AI-powered migration. Built exclusively for founders. https://t.co/rJ9xqHvyGq
At @Techweek_ in New York, our AI Visionaries Forum brought together founders, VCs, and execs who are building the next generation of AI systems. These systems are ones that learn and improve in production, not just in the lab. We heard from the founders of @edra_ai and @ElorianAI on how they're scaling these systems, and got the investor take on why speed of improvement is becoming the defining metric for success. Amrita Sarkar, Frontier AI Startup GTM Leader at AWS, shares what's shifting and why it matters.
Made in Seoul with AWS: Sunbin You & Dogyun Kim, DALPHA. Partnering with AWS from the beginning helped DALPHA move fast without worrying about infrastructure, the startup’s founding team explain. They share how building in Seoul, surrounded by talent, “gives us energy” and offers an environment that “pushes us to keep going.”
In just a year, @GeminiApp users have more than doubled, surpassing 900 million. #GoogleIO https://t.co/SZoixMPv0d
Gemini 3.5 Flash ARC-AGI (Verified) ARC-AGI-2: - High: 72.1%, $0.85 - Minimal: 8.9%, $0.11 ARC-AGI-1: - High: 92.5%, $0.42 - Minimal: 48.8%, $0.06 Gemini 3.5 Flash is on par with GPT-5.5 (Medium) on ARC-AGI https://t.co/UqzO7KbVZ7
@SouthernWintrs Here’s me and my lab. I’m an area chair for the most prestigious publication venue in my field. I think he’s completely correct and that behaving otherwise is academic misconduct. I would fire someone for doing it habitually. https://t.co/7Ao3ZqSQzt
Details about PyTorch Conference North America at: https://t.co/Qh00NZiMt7 Note: the CFP for speaking submissions is currently open!
And details about the joint KubeCon + CloudNativeCon + OpenInfra Summit + PyTorch Conference China 2026 are at: https://t.co/5AGUljH3EN
The clock is ticking! ⏰ Only 1 week remains to submit your proposal for #PyTorchCon North America in San Jose (Oct 20-21). Get your sessions in by June 7. Submit: https://t.co/hLlKK7WxLD https://t.co/U8ZJjeDTv5
PyTorch Foundation member @NVIDIA just adopted OpenMDW-1.1 across four of its biggest open model families. OpenMDW is the Linux Foundation's permissive licensing framework — originally launched in 2025 with the PyTorch Foundation — purpose-built for AI model distributions instead of being retrofitted from traditional software licenses. By replacing fragmented, restrictive AI licenses with one unified legal framework covering models, weights, code, documentation, and data, OpenMDW gives developers clear rights to train, modify, contribute, redistribute, and deploy. 🔗 Learn more: https://t.co/ljOff6ozDb
AI is moving from software into the physical economy. Amazon’s new warehouse robots show that the next productivity wave will not only happen on screens, but also in logistics, factories and supply chains. The uncomfortable question is how fast companies can retrain workers while automation keeps getting smarter.
Your digital twin may become your most productive employee. What started as AI helping people write is evolving into AI that can communicate, answer questions and eventually represent experts, founders and executives at scale. The next competitive advantage may not be having AI, but having an AI that thinks like you.
AI is no longer just a technology race. It is becoming a wealth-distribution debate. If Americans receive equity stakes in AI companies, it would be a historic shift from taxing innovation after the fact to sharing ownership from the start. The big question is whether this builds public trust in AI or opens the door to political control of the sector.
The future of finance is invisible. Agentic AI, embedded finance and intelligent infrastructure are transforming finance from a destination into an always-on layer woven into everyday life. The biggest winners may not be the brands consumers see, but the companies building the rails, orchestration and trust behind the scenes.
AI’s real power is not only in productivity. It is also in protection. Using AI to detect illegal marine wildlife trafficking shows how the technology can help defend biodiversity, borders and fragile ecosystems. The best AI use cases may be the ones that make invisible crimes harder to hide.
https://t.co/csgtuQ5qSh
The AI boom is no longer defined by breakthroughs. It is defined by capital. Trillions are flowing into AI infrastructure while companies race to prove that the returns can justify the spending. The next phase of the AI race may not be about who builds the best models, but who turns them into sustainable profits.
https://t.co/M1BfgRQsI7
The AI race is moving from private funding rounds to public markets. StepFun’s planned Hong Kong IPO shows how Chinese AI companies are trying to capture investor appetite while the window is open. The next test for AI will not only be technical performance. It will be whether public markets believe the valuations.
Periodic reminder to look at the data https://t.co/H5XlmlQLUH
Today we're announcing our Series C funding: $355M at a $4.65B valuation, led by some great investors @generalcatalyst and @Redpoint. We've had insane growth in the last year, but we're still very early. So proud of the team and what we have built so far! https://t.co/crr5Oi6L5j
i'm restarting my blog! i want to kickstart productive conversations around: what should AI agents look like for hard, subjective knowledge work? a lot of agent setups work well when tasks are objective and easy to verify. but many workflows (e.g., qualitative analysis, strategy, sensemaking) are messy and interpretive. as a first post, i explore different ways of doing agent-assisted qualitative analysis on tweets, with varying levels of human feedback/intervention. tldr: they all kinda sucked. turns out it’s hard to: (a) stop agents from converging too quickly on shallow interpretations (b) get agents to adapt to preferences that emerge gradually across many turns (i.e., evolving context) (c) capture human judgment without making humans fatigued
It's a little early to say, but I *think* the glowy cells here in this tiny duckweed are transiently expressing some genes I had made, coding for a fluorescent reporter + miraculin! Replication + more controls this weekend. https://t.co/hImksQzdKM
The pair of adjacent cells is a good sign - makes me hope that was a single cell that got transformed then duplicated. @NikoMcCarty duckweed downside, these cells are so damn small 😂 model organisms should have onion-sized chonkers lol. https://t.co/5A68D5s7qz
@NikoMcCarty This DNA is actually intended for agrobacterium transformation, which will hopefully give me completely transformed fronds. But the gene gun test is nice and quick while I'm blocked on getting some antibiotics (just had another order cancelled because I'm just an individual) https://t.co/7OyCHiqiHe
@NikoMcCarty One single algae cell might have caught a stray! One reason I'm so careful to make sure everything potentially transformed ends up in a vat of bleach :) but I might have to try for more algae transform, that would be so cool! https://t.co/D8Yz3wyLHy
