The bright and sunny energy of LA was found at AWS Summit Los Angeles! The air was buzzing with founders comparing notes, builders sharing real-world lessons, and investors and operators exploring the latest innovations across AI and cloud. This is what's possible when the startup community comes together.

Regretting that commit? Don't panic! Here's how to fix it. Check out our latest GitHub for Beginners episode, answering some of your most commonly asked questions 👇 https://t.co/p87G5DFU0a https://t.co/4KpK64eWG2

This video by @jbhuang0604 is a compact but very informative dive into the progress of self-supervised learning over the past few decades. from IMAX in 1992 covering methods like MoCo, SimCLR, DINO, BYOL, MAE all the way up to LeJEPA in 2025 Highly recommend watching! https://t.co/4EkdfHHxWV

A tremendously important and short (2mins) read: https://t.co/jiBlax5YoW https://t.co/vNnQRZwHtI

A tremendously important and short (2mins) read: https://t.co/jiBlax5YoW https://t.co/vNnQRZwHtI

Sakana AI、初の商用プロダクト「Sakana Marlin」を提供開始 リリースブログ：https://t.co/KeTds8RtTR https://t.co/iI1GR2rKEG

Sakana AI、初の商用プロダクト「Sakana Marlin」を提供開始 リリースブログ：https://t.co/KeTds8RtTR https://t.co/iI1GR2rKEG

@segyges Sorry https://t.co/dMyS1sFo6j

🚨 BREAKING: Trump's new peace deal with Iran includes a $340 billion payment by the United States to Iran. https://t.co/pED8KHAKwP

“But even before we have the details, it is clear that Trump has failed to achieve every one of the goals he put forward for this war of choice, and now he is determined to sign, seal, and deliver America’s capitulation as quickly as possible” https://t.co/pCukqG35Gn

“You are not a free citizen participating in a noble democracy. You are simply tax-generating livestock, permitted to occasionally vote for which butcher holds the knife.” -Lysander Spooner https://t.co/UVD2F00o1g

@cobracommanduhr @axios see this letter and signatories: Open Letter on Transparent AI Cyber Protections https://t.co/60GBgnRIrM

What started as a technical issue has quickly become a policy issue. Anthropic is now sending staff to Washington to resolve restrictions on its most advanced AI models, highlighting how closely AI development and national security have become intertwined. A few years ago, AI companies were competing for users. Today, they are increasingly negotiating with governments.

Important news via @axios “Prominent cybersecurity leaders — including CISOs, security researchers and executives at Adobe, Zoom and Sophos — are urging the Trump administration to reverse restrictions on Anthropic's most advanced AI models, arguing the move hurts cyber defenders more than attackers.”

Github (uploaded by Claude 4.8 Opus, which also added a text size slider, I didn't let Opus touch the somewhat odd prose that was typical of Fable 5): https://t.co/h8eb4h5BbT

Gory details of KPMG’s ‘Redefining Excellence’ paper, which is atrocious AI slop. 40 of 45 citations can’t be corroborated. It’s a historic own goal. https://t.co/pAKOPdwOza https://t.co/FcuP9ng0KA

That one Codex Thursday when we shipped codex remote control with m. https://t.co/2lYWwYonV9

+2 more

I'm so terrible at discord I'm creating a slack community for https://t.co/on6yrIlqhC :) 2021 vibes

I'm so terrible at discord I'm creating a slack community for https://t.co/on6yrIlqhC :) 2021 vibes

Sakana AI、初の商用サービスはリサーチ特化　「Deep Research」との違いは？　後発で“ベンチマークも追わない”ワケ https://t.co/ZnpqDXGMfs

Sakana AI、初の商用サービスはリサーチ特化　「Deep Research」との違いは？　後発で“ベンチマークも追わない”ワケ https://t.co/ZnpqDXGMfs

🫡 https://t.co/7cM1lzz1Fn

Starlink ad on the Octagon at today’s UFC Freedom 250 event at the White House. https://t.co/0JgdLU3zE1

🐠 Sakana Marlin リリース 🐠 本日、Sakana AIは、初の商用プロダクトとして、ビジネス向けの自律型リサーチアシスタント「Sakana Marlin」を提供開始しました。 詳細はこちら：https://t.co/ZWsYW91LGu ブログ：https://t.co/leVp2dPxgL

+1 more

Should we be worried about labour displacement at the hands of AI? "Short term, yes. Long term, no. In the near term, AI is creating real disruption, with large layoffs and thousands of people being displaced from existing jobs. Over the long run, there are far more problems to solve than people currently working on them. If AI makes engineers more abundant and productive, we can redirect that talent toward solving problems that were previously uneconomic or impossible to tackle." @matanSF Do you agree and what does no one consider that everyone should with regards to future of labour @pmarca @MattEvantic @shaunmmaguire @vkhosla

laser weeding in broccoli. real-time AI decision making at the edge. whistles and purrs. no chemicals. from our friends @carbon_robotics. https://t.co/dF84n71GMl

check out my /ultragoal skill https://t.co/aPya2D55zZ

Was cool to be able to celebrate the @SpaceX IPO with a few folks from the amazing @Starlink team. Ultimately SpaceX is about building cool shit with your friends, now that’s something that will keep you going … https://t.co/OhzEP22AcY

MIT, Stanford, New York Univ, Princeton paper says AI can make people feel more efficient even when they are not actually becoming much more efficient. that people often use AI for simple tasks because it feels like it saves time and effort, but the measured benefit is often tiny, missing, or even negative. The biggest point is the feedback loop: once people use AI, they become more likely to use it again, even for easy tasks where doing it themselves would often be just as fast or faster. i.e. AI dependence can grow from a mistaken feeling of convenience, not just from real productivity gains. Across three preregistered studies with 2,691 participants, people used AI for basic arithmetic, spelling, recall, and short rewriting at higher rates than they predicted, especially on easy tasks. They also expected AI to save 55.7 seconds on average, when the measured saving was only 7.5 seconds. For simple work, the hidden cost is not intelligence but interface friction: writing the prompt, waiting, reading, checking, and deciding whether the answer is acceptable. Once that loop begins, it can feel like effort has been outsourced, even when effort has only been rearranged. Here’s the key part: the study suggests that AI use can train its own justification. After using AI on just two tasks, participants became more likely to use it again, even when independent completion was faster. The danger is not dramatic dependence, but quiet recalibration. A person who asks AI for a trivial answer today may not become less capable tomorrow, but they may become less accurate at judging when their own mind is already the faster tool. ---- Paper Link – arxiv. org/abs/2605.22687 Paper Title: "The efficiency-gain illusion: People underestimate the rate of AI use and overestimate its benefits on simple tasks"