Activity
4K followers
-
Elan Dekel shared thisI built a small experiment: an AI focus group. Enter a URL or a product concept and get feedback from an AI consumer panel, an AI startup investor panel, or an AI UX design panel. The panel feedback is actually pretty interesting! It'll be interesting to see if AI focus groups can make early product and prototype testing faster. Try it here and let me know how it works for you: aiagentpanels.vercel.app
-
Elan Dekel shared thisIt took some time, but now artists are starting to learn how to use AI tools. From AI slop, to true creativity powered by AI magic. Lovely video by Anton ChernykhElan Dekel shared thisThe sweetest compliment I ever got as a screenwriter was: “Give him the phone book and he’ll find a story in it.” Well, the phone book is next. For now, I’ve turned Jean-Léon Gérôme’s classic "The Duel After the Masquerade" into a story. What once required an artist to navigate miles of producers now takes only a few subscriptions. Good times ahead. Nano Banana Pro/ GPT Image Model 2/ Kling/ Adobe Premiere/ my brother with a touch of CG
-
Elan Dekel shared thisI love this story. I worked with Lazer at his previous company when I was at Pinecone and was very impressed. I was deep into OpenClaw in January this year (yup I bought a Mac Mini and went to the first ever Clawcon) and was blown away when I saw what Lazer and Gavriel were doing with NanoClaw - solving the key security concerns holding back adoption. Very excited to see them continue crushing it!Elan Dekel shared thisToday, we announced $12 million in funding, and I was referred to by Fortune as "Lazer, the elder". Tens of thousands of people are using NanoClaw and getting tremendous value, and many have reached out requesting AI agents for every member of their team. NanoCo (creators of NanoClaw) delivers a professional assistant to every employee within Slack, Teams, and other tools they already use. It learns the employee's role, adapts to their style, and builds a persistent understanding of them so that it can get real work done. It's hard to imagine this journey began only three months ago, and I can't wait for what's next. Thank you to our amazing team, our investors - including Valley Capital Partners, Vercel, Docker, Inc and monday.com - and users 💪
-
Elan Dekel reposted thisElan Dekel reposted this🚀 New from AGI Lab: Nova Act Quick Deploy Studio One click, and you've got a full browser-automation studio running in your own AWS account. No managed service in front of it, no data leaving your environment. Try it: https://lnkd.in/g5Vh3i_a What you get: ✦ Author workflows in natural language ✦ Watch the agent drive a live Chromium session in real time ✦ Schedule any workflow on a cron or rate ✦ Full run history with screenshots and traces ✦ Everything (UI, auth, runtime, storage) provisioned from a single CloudFormation template It's powered by Nova Act, so browser sessions and act steps run under your account's IAM permissions. Bring your own AWS, own your data, ship automations in an afternoon. Thanks to Fede G. and Ryan Petersen for building this! 🙌
-
Elan Dekel shared thisAn interesting read if you’d like to learn more about how computer-use models are trained. It's great to see some of the innovative work coming out of our lab.Elan Dekel shared thisTraining AI agents to click, type, and navigate real websites requires more than just a great model. Amazon's AGI Lab shares a practical recipe for scaling reinforcement learning into a reliable engine for computer-use agents. https://amzn.to/4vA60CD
-
Elan Dekel reposted thisElan Dekel reposted thisAfter 27 years at Amazon, I figured it was time to end my quiet period on LinkedIn. Reading Andy's shareholder letter was all the nudge I needed. I never thought I’d be at the same company for 27 years… it just sneaks up on you when you’re having fun building things with smart people. One thing I love most about Amazon is our willingness to think big and be patient. When I joined, we were a bookstore with ambitions to deliver everything to everyone. That vision consumed my first decade. Then I helped launch EC2 – the start of a 20-year journey building AWS services and expanding our global infrastructure. A little over a decade ago, I started working with the Annapurna team on our chips business. When Andy wrote about it last week, I stopped mid-read – seeing it laid out so clearly reminded me of how far the team has come. Building our own chips wasn't an obvious call. There was real skepticism, and good reasons to avoid the complexity. But we believed that controlling things down to the silicon meant we could optimize AWS in ways that truly served our customers. That bet paid off. Every EC2 instance runs on our Nitro chip. Graviton is used by 98% of our top 1,000 EC2 customers. Trainium2 is essentially sold out. Trainium3 just started shipping and is nearly fully subscribed. None of this happened overnight, and none of it happened without a team willing to be patient and a little misunderstood along the way. I've watched this cycle play out often at Amazon. Bets that looked wrong, then right, then inevitable. It’s made possible by a culture that gives long-term thinking room to breathe while maintaining the urgency to move forward. Three months ago, I stepped into a new role bringing together our foundational AI models, custom silicon, and quantum computing. People often ask why bring these together, and why now. The question was never if but when. When model scientists know where infrastructure is heading, they explore techniques that wouldn’t otherwise be viable. When chip designers know where model architectures are going, they optimize differently. When you build models on your own chips, you train at greater scale and experiment faster. And as quantum matures, it will benefit from everything we've learned scaling the other two – especially our hard-won expertise designing some of the world's most complex chips. The industry is moving fast and there’s a lot of noise. But what I keep coming back to is how early we are in this massive transformation. Optimal AI infrastructure is still being invented by us and others. And today's models are genuinely amazing, but in a few years, we'll look back at them like dial-up internet... slow and expensive. We're working on problems that don't have obvious answers yet – and some of what we're doing will look strange for a while. At Amazon, that's usually a good sign. Read Andy's letter here: https://lnkd.in/gMFAYKHT
-
Elan Dekel reposted thisElan Dekel reposted thisNew roles just opened on my team that are pretty great! We support a newer org that brings together AI, chips, and quantum - and these two roles specifically sit on the chips side of the business, which is growing incredibly fast with no shortage of interesting stories to tell. The team itself is a genuine highlight. Smart, a little kooky, and a lot of fun 😊 Roles are open in Seattle, SF, or NYC. Message me if you're curious - happy to chat! PR Specialist: https://lnkd.in/eJWwA-Fd PR Manager: https://lnkd.in/eypAZ-J8
-
Elan Dekel reposted thisElan Dekel reposted thisWe're releasing our fastest and most accurate web search API for real time news to the public. And giving you $100 credits to try it. Why? We built our web search API from scratch. Our own crawler, our own indexing pipeline, our own retrieval models. The results were worth it. Every day we benchmark ourselves against every major search API provider. On April 11th, our API produced high quality results for US news 7x faster than everyone else (benchmark data below). I wrote about why we built everything from scratch, started with news, and what we hope to accomplish from here. I’m thankful to the team for choosing the hard path, and proving it's the right one. Are you building agents that need real-time web knowledge? Try us. Links in the comments.
-
Elan Dekel reposted thisElan Dekel reposted thisThe MCP Dev Summit in NYC last week was an amazing experience. Huge thanks to the Agentic AI Foundation for putting it together and bringing this community of builders together! 🎉 We launched Amazon Nova Act MCP at the event, an MCP server that lets your AI coding assistant automate UI workflows through natural language. Need your agent to extract data from a website, fill out forms, or run UI tests? Now it can. Behind the scenes, Nova Act handles the hard parts: picking the right browser, dealing with bot detection via Web Bot Auth (emerging IETF standard), running parallel sessions. So your agent just says what it wants done. Works with Claude Code, Kiro, Cursor, or any MCP client. ➡️ One command to get started: uvx amazon-nova-act-mcp --configure Try it out and let us know what you think! https://lnkd.in/ean3hXDJ Kudos to Emile Baizel and Yilin Zhu from Amazon AGI Lab for creating the MCP server. #MCP #AIAgents #NovaAct #DeveloperTools
-
Elan Dekel liked thisElan Dekel liked thisPersonal headline: I’m becoming an insurance broker. After two amazing decades in consumer tech, that may seem like a plot twist. But it’s less surprising for anyone who has been following my journey since my hometown burned down a year and a half ago. Our work after the fires in LA showed us how critical it is to have the right insurance coverage for your needs. But many folks have coverages that were and are out of alignment. The stats in LA are staggering. The non-profit Department of Angels reports that as many as 75% of fire survivors were underinsured, with as many as 43% going into debt as a result. This doesn't need to be the case. In the same way that a wealth manager shapes upside, at Bevel we believe families now need a real risk manager, a true quarterback to help them understand how to protect the assets and the life that they've built. We've raised capital from some amazing investors (SHAKTI, Resolute Ventures, Bessemer Venture Partners) and incredible individual investors too. We’re partnering with excellent insurance companies and leaders to build what we believe is the future of private risk – helping families stay safer in an ever more complicated world. Please also know that if you're one of the thousands of people who has been on the journey with us using our AI-powered tools to get your insurance claims paid out, we are dedicated to continuing that work for free. We'd love to have you join us on the journey. You can read my founder’s letter or our press release following the links in the comments. You can follow Bevel for information about insurance.
-
Elan Dekel liked thisElan Dekel liked thisA few weeks back I said I'd been cold-emailing people I admire and asking to interview them about their careers. Antje Barth said yes, which I'm still a little amazed by. She doesn't need an introduction from me, she's a champion of this industry. She went from network engineering to developer advocacy at AWS, wrote a couple of books along the way, and she's now at Amazon AGI Labs. She's basically had four careers, and each one started the same way, by spotting where the frontier had moved and going there. Networking, then big data, then cloud, and now agentic AI, which she says is the hardest one so far because none of the patterns exist yet. What stuck with me is that none of it turned out the way she pictured it, and she thinks that's the best part. The plan was never the point. She also said the safe move is almost always the one you regret. Something that I would like to work on. Thanks Antje, for doing this and for not sending back the tidy version. It's up now, unedited. 𝗭𝗲𝗿𝗼 𝗧𝗼𝗸𝗲𝗻. 𝗔𝗹𝗹 𝗛𝘂𝗺𝗮𝗻. https://lnkd.in/dWEFgeWu
-
Elan Dekel liked thisElan Dekel liked thisPersistence is usually considered a virtue, but sometimes the wisest thing we can do is give up. In our new article, Aly Lidayan, Anne Collins, and I investigate why people persist with goals even when better alternatives become available. One possibility is that people avoid switching goals because of the cognitive cost of reconfiguring how they represent the task at hand. Across two experiments, we found that these costs do influence goal persistence and which alternatives people choose, but explain only a small part of the overall persistence bias. Understanding what else drives stubborn persistence is an exciting question for future work. Link to the full CogSci paper in the comments! Thanks to the University of California, Berkeley Institute of Cognitive and Brain Sciences, Psychology Department, Helen Wills Neuroscience Institute at UC Berkeley, Center for Human-Compatible AI, Berkeley Artificial Intelligence Research, and Hertz Foundation for sponsoring this work.
-
Elan Dekel liked thisElan Dekel liked thisThis is a post for Israeli entrepreneurs who are starting a company and are debating between US and Israel incorporation. I generally don't do "thought leadership" and "advice to entrepreneurs" posts, but I recently saw too many cases of founders who made a decision only to regret it almost immediately, and the consequences can easily be millions of dollars in personal tax per founder. Or worse — in some cases, admittedly more rare, the decision may jeopardize the company. The bottom line is that I've never met a founder who regretted a US incorporation, and I often meet founders who regret an Israeli incorporation. For some reason, some lawyers and accountants in Israel advise incorporating in Israel. There may be some advantages, but they are all pretty small. And the cynic in me thinks that those accountants and lawyers might benefit from the Israeli incorporation, as their services will be needed for longer. In contrast, there are at least two big reasons to incorporate in the US: 1. QSBS. This is a huge personal tax benefit for shareholders in a US startup. In theory it's limited to a cap of $15M exemption, but it can actually be stacked so the effective cap can be much higher — and it is only available to US companies. When you start a company in Israel, there is a very good chance that one of the founders will at some point relocate to the US and become a US tax person. Or they may already be US citizens. And in any event, they will likely have US investors and US employees. The decision to incorporate in Israel robs the founders, the investors, and in some cases even early employees who choose to exercise their options early, of this benefit for no good reason. 2. In some areas there is a real advantage in being a US company — for example, if there is a chance you'd be selling to a US government entity. I recently saw a company that could very well go out of business because of complications that could have been avoided had they been US incorporated. Now, admittedly, unlike the first reason, this one is probably only relevant in fewer than 5% of cases, but the consequences may be more severe. In addition to these two big reasons, being a US company, typically Delaware, just simplifies things from a legal and corporate governance perspective on a number of topics. It is important to note that even when the company is US-based, Israel still benefits from taxes, both on an ongoing basis and, of course, upon exit (e.g., the Wiz exit reportedly generated several billions in tax revenue for Israel). In fact, this is the reason Israel makes it very easy to change the incorporation to the US (הפיכת שרוול), at least early in the life of the company. But the problem is that this does NOT help the founders, as QSBS only applies to shareholders who originally received their shares in a US company. Disclosure: I personally benefit when I invest in a US company vs. an Israeli one, so I am financially incentivized to give this advice.
-
Elan Dekel reacted on thisElan Dekel reacted on thisI recently asked a senior leader at a very well-known startup with ~1,000 employees: “With how fast the innovation cycle is and how intense the competition is, how does the team collaborate? What’s the collaboration culture like?” I got a genuinely puzzled look back. The kind of look that says: why are you asking me this? Then: “We’re a startup. We just solve problems. Whatever it takes to solve problems.” 😂 I was expecting some version of how teams coordinate, how decisions get made, how things work across functions. Apparently the operating model is: Problem exists. Solve problem. Next problem. Honestly, maybe I’ve spent too long at big companies.
-
Elan Dekel reacted on thisElan Dekel reacted on thisExcited to share that I started at Anthropic this week, joining the RL Human Data team. I'm feeling incredibly lucky to work alongside such brilliant people on some of the most important challenges in AI. Looking forward to what’s ahead!
-
Elan Dekel liked thisElan Dekel liked thisAfter 16 unforgettable years, the time has come for me to say goodbye to Google. 👋 When I joined in 2010 through the acquisition of Like.com, it felt like stepping into a different planet: moving from a scrappy startup where I could run the whole tech stack on my laptop to a “huge” company of 10K employees and a massive codebase. Fast forward 16 years, and I’ve had the privilege of coming full circle—helping Google acquire and integrate other amazing companies across the globe. Looking back, it’s hard to capture everything in a single post, but a few chapters will always stay with me: Sixteen years of AI in Google, one classifier at a time. A first step was teaching a model to recognize "red strappy sandals" 👠 or recognizing A-line dresses 👗. I could not have seen in 2010 that I would end up building enterprise LLMs and voicebots 16 years later. The Google Shopping Transformation: From the early days of free product search (Froogle anyone?) to building a core pillar of Google’s ads engine, I led engineering across Shopping, starting with visual search to building the data warehouse, analytics, and pricing insights that quietly sit behind a lot of your online shopping today. I took on the ambitious mission of cataloging "all the products in the world at the best prices", applied AI/ML/CV for scaling to over 60B+ products. A year that changed how I see search. I led Lens Quality & Experience, helping scale it from 100M users to a billion and shipping Multisearch. This was one of the more humbling stretches of my career, watching visual search go from "neat trick" to something people reach for without thinking. Cloud AI and the wildest chapter yet. It started with building POCs and developing cutting edge AI research for enterprises 4 years ago, and evolved continuously with the wide adoption of LLMs. Over the years I owned the vision and GTM for Google's specialized enterprise AI (speech, translation, document understanding, security and Responsible AI). Watching enterprise AI adoption go from slow build to explosion, as an insider, was a genuine privilege. Mountain View to Zürich. Starting at HQ, finishing in Switzerland, I have worked alongside genuinely brilliant people across US and Europe. As cliché as it sounds, the absolute best part of Google has always been the people. To my managers, leaders, and mentors over the years: thank you for your guidance, wisdom, and trust. To all the extraordinary teams I’ve had the privilege of leading: thank you for your dedication, brilliance, and for constantly inspiring me. And to all my former colleagues, many of whom are now lifelong friends, I owe you my deepest gratitude. You made every challenge fun and every success meaningful. What's next? I'm taking a real breather to recharge, and I'll be advising a handful of companies on building enterprise AI adoption. If you're working on something interesting in that space, or just want to catch up over coffee (ZH or virtual), ping me.
Experience & Education
Other similar profiles
Explore more posts
-
Jay Mehta
Tech’s parenting debate just entered a new era. Meta’s AI chief Alexandr Wang recently said he plans to delay having children until brain–computer interfaces like Neuralink are “very advanced,” arguing that early childhood neuroplasticity makes kids uniquely suited to grow up as native users of direct AI–brain links. That vision is extreme, but it perfectly captures how family planning is shifting in a world where technology isn’t just changing our jobs, it’s redefining what it means to raise the next generation. On one side, critics see this as dystopian: children reduced to “human capital upgrades,” plus unresolved risks around safety, consent, inequality, and long‑term mental health. On the other, some in tech argue that delaying parenthood until certain technologies, financial independence, or climate risks are clearer is a rational portfolio choice on life itself, similar to how more professionals are using fertility tech, remote work, and long-term career planning before starting families. For innovators and investors, this isn’t just gossip about one founder. It’s a signal: consumer expectations, bioethics, regulation, and markets around fertility, neurotech, and human performance are going to be massive thematic arenas. What do you think: is “tech‑timed parenthood” visionary risk management or a step too far? #Innovation #BioTech #Neurotechnology
-
Dan Gonzales
Ali Dastjerdi CEO of Raylu shared his story on Startup Mindsets and the complex world of AI and how to work with LLMs. It was quite a fascinating conversation: listen on Spotify for the full episode: https://lnkd.in/gbvwirUd How are you using LLMs and has it been to your benefit? LMK in the comments! #StartupLife #Founders #Entrepreneurship #BuildInPublic #StartupJourney #Startuplife #VentureCapital #SeriesA #SaaS #GrowthMindset #BusinessStrategy
7 Comments
Explore top content on LinkedIn
Find curated posts and insights for relevant topics all in one place.
View top content