Connor Hayes

🚀 We've open-sourced TimesFM – a decoder-only foundation model for time-series forecasting 📈 Key Highlights: ✅ Large-Scale Training Data: Trained on a corpus of over 100 billion real-world time points, including Google Trends and Wikipedia pageviews. ✅ Zero-Shot Forecasting: Offers high-quality forecasts even on unseen datasets, without requiring fine-tuning or additional training. ✅ Outstanding Benchmark Performance: Achieved competitive results across a comprehensive suite of benchmarks. 🤗 Explore the model on Hugging Face and GitHub. I'm encouraged by the early feedback and look forward to seeing how #TimesFM will be used in real-world applications for forecasting and anomaly detection. 🙏 A special thanks to the talented research team behind this achievement: Abhimanyu Das Rajat Sen Yichen Zhou Weihao Kong

42

Marc Rebillet is shooting robes out of a cannon at the Google developer conference while demoing the Generative AI DJ features. Product launches do not need to be boring. Tightly scripted, highly produced, pre-made product content is OUT. Next level stuff at Google IO (#googleio). And this is just to warm up the crowd. After this and yesterday's very live (warts and all) approach to OpenAI's product launches.. I think its clear that authenticity and the "real" is easier for consumers to connect to than the tightly scripted (show-no-flaws) approach that Apple took with the recent iPad launch event. To my eye, I think everyone should strive to show more humanity and be open to show flaws in these comms. Especially when the world is shifting to a place where it's harder to tell generated AI content from real, human-produced content. #productmarketing

5

6 Comments

Isar Meitis has does a great job creating content covering not only the what of AI use cases but spending time diving into the how. I had a great time chatting with him for this episode that went into details on one of the most entry-level AI + Automation use cases I share and then into some of the most advanced ones. We also got a chance to talk about Zapier's first agentic tool, Central (Which you can try for free today!) All the links in the comments here. 👇

21

2 Comments

Hey Friends, Today we're revealing the biggest launch in the history of Superhuman: reimagined for teams!! I would greatly appreciate your support if you could upvote us on Product Hunt https://lnkd.in/gX_b3PVi (if you haven't already, update your Superhuman version to get access the game-changing features)

44

5 Comments

I am so proud of the work that is being delivered by our team in Search that was showcased at I/O this week. Highlights were... ⚡Plan Ahead with Generative AI in Search ⚡ Thanks to a new Gemini model customized for Google Search, we are able to do SO MUCH more. With planning capabilities *directly* in Search, you can get help creating plans for whatever you need, starting with meals and holidays. Search for something like “create a 3 day meal plan for a group that’s easy to prepare,” and you’ll get a starting point with a wide range of recipes from across the web. And if you don’t like the suggestions, Search will continue working on your behalf. So if you want to swap the dinner suggestion for a vegetarian option, just like that, Search will customize your meal plan. On top of this, there are lots more updates in Search so — for the curious  — have a read of the blog by Head of Search Elizabeth Reid below 🔽 https://lnkd.in/en6_7fvX #GoogleIO #GenAI #GoogleSearch

177

7 Comments

🧠🧠🧠 Let's not forget that daily model updates does not mean marketers need to leverage their MMM every single day. It means that our data is ready when YOU need it, not when we feel like running the model. Does that cost more for us? Sure it does. But maybe doing things just a bit differently is what our industry needs. Reach out if you have questions 🤙🏼

4

Excellent study from Damian Rollison and the SOCi, Inc. team which seems to confirm my longtime hypothesis around citations (https://lnkd.in/gxZQwdx3): the only citations that are valuable are the ones that rank in Google for the keywords you want to be known for. https://lnkd.in/g49DMSeN

39

11 Comments

🚀🚀 Exciting news! 🚀 🚀 I left Google two months ago to start a new company - Aligned Labs - with Tyler Garvin At Google, I had the unique opportunity to work at the intersection of product, design, and strategy for our large language models, like Gemini. This experience made one thing crystal clear: data makes the difference. As we push the boundaries of AI capabilities, we're rapidly depleting our sources of unique, high-quality data. This isn't just a speed bump—it's a potential roadblock for the entire industry. At Aligned, we're not just collecting data—we're curating a first-of-its-kind team of the world’s leading domain experts and educators. Our mission is to create top-tier, unique datasets for training large language models. Our goal? To ensure AI outputs align with our expectations of highly intelligent thought partners. We're crafting datasets for every stage: pre-training, instruction-tuning, fine-tuning, and evaluation to empower our clients to unlock the full potential of their models and applications. I'm dedicating the rest of my career to this for three critical reasons: - The data crisis is real and imminent. Without a solution, LLM performance will plateau, stalling AI progress. - As models advance, they require increasingly specialized data. We need to tap into expert knowledge across various domains, but no one has cracked the code on how to scalably attract and retain this talent. Companies like Scale AI have done an amazing job of building a workforce of cheap educated workers but what you really need for the next level of performance are domain experts that can train and impart their expertise and knowledge into these models. - Frontier model builders like Google, OpenAI, Meta, and Anthropic understand this problem and are making massive investments, there's a huge opportunity in specialized data. That opportunity goes beyond the Google’s and OpenAI’s of the world. High-quality, unique datasets can dramatically improve AI applications in specific industries, making them perform better and more cost-effectively. We're here to unlock this potential, helping companies of all sizes harness the power of tailored, premium data to build superior models and applications for their specific needs. We're incredibly excited about the potential of our work to dramatically improve AI models and help companies across industries harness the power of truly intelligent AI. Are you passionate about the future of AI? Let's connect! Whether you're a potential client, an AI enthusiast, or a domain expert interested in contributing to the next wave of AI advancement, I'd love to hear from you. Head over to https://www.alignedhq.ai to connect! #AIStartup #MachineLearning #DataQuality #FutureOfAI #AlignedLabs

281

75 Comments

Breaking down Amazon's 'Just Walk Out' AI Pause Why did the tech giant stop its revolutionary "Just Walk Out" technology? CMSWire's Pierre DeBois does a deep dive into what this means for the future of #AI in #customerexperience. So what went wrong with Amazon’s cashier-less tech? Despite its ambitious vision, Amazon faced significant hurdles. Understanding these can help guide your own AI strategies. As AI increasingly shapes #customerjourneys, it’s vital to implement strategies that don't compromise on trust. What strategies are you employing in your AI deployments? Want to learn more about Amazon’s 'Just Walk Out' pause. Read the full article here: https://lnkd.in/eGtE5eEt #CustomerTrust #AIEthics #Amazon

9

After 10+ years working with early-stage startups it’s fair to say I’ve seen I’ve seen some things (& some things I will never unsee), learned some things, earned some stripes, some scars, and some measure of expertise (one hopes) helping more than a dozen startups find product-market fit. While I’ve enjoyed year-long runways at well-funded startups that were eventually acquired, or went public, or very publicly imploded—what really taught me the real meaning of Lean startup was working with self-funded founders who had ten compelling concepts but resources to build one. When you don’t have a year to be wrong; you learn to get it right very fast. If you are a founder facing your own set of high stakes decisions & tradeoffs; we can help. For free. You can book a 1 hour chat with us to talk through all the things racing through your head. We know what’s involved in getting something tested, validated, and built; this is what we do every day all day. We might not have all the answers but there is a good chance we can give you some clarity, some calm, some options. Call us at 1.844.4FLAGRANT or DM me. And if you want to get a sense of what we are like as hoomans- check out our Delightful Reveal series. It’s all the things we have been reading, watching, listening to, & enjoying. https://lnkd.in/gfrz9tQS #startups #entrepreneurship #socialentrepreneurs #innovation

30

5 Comments

I created this Reforge guide on 𝐡𝐨𝐰 𝐭𝐨 𝐰𝐫𝐢𝐭𝐞 𝐚𝐧 𝐀𝐈 𝐩𝐫𝐨𝐝𝐮𝐜𝐭 𝐬𝐩𝐞𝐜, because I’ve actually never written a spec this way (!). Instead, while building AI products, I’ve written fragments of problem / root cause / solution, chatbot conversation examples, logic flowcharts, QA docs filled with screenshots, all while thinking there must be a better way to do this. I’ve also been asked many questions by PMs who are figuring out how to build AI features. How do you figure out what an LLM can do? Does it have to be a chatbot? What’s the same as or different from building a non-AI feature? Writing this guide helped me crystallize a few learnings: ✨ If you’ve been given a hammer and need to look for nails, go find good nails. ✨ LLMs are non-deterministic, which means you need to articulate a range of behaviors and mitigate risks. ✨ Thoughtful UX can make all the difference in whether your users get value out of AI. This guide represents how I’d like to approach my next AI feature. I hope you’ll find it useful. Many thanks to Wendy Lei Akshita Ganesh Mahesh Ram John Beckmann Madison Holbrook Helen Lee Kupp for their invaluable input! https://lnkd.in/gim6fHdd

285

17 Comments

Love how Steve Jobs talks about conflict resolution. For many years at Google, the notion of "disagree and commit" didn't exist. When there were differences of opinion, we duked it out (with words and vigorous debate). Looking back, it seems to me that if the culture has the following attributes, then you have the capacity to reach consensus: - Humility: being open to "maybe I don't have the best answer" - Curiosity and Respect for peers: "I see some merits in my peer's ideas" - Drive to find the optimal path: "let's huddle till we sort this out" Very often, when evaluating two options, the vigorous debate would lead to a new third option that was better than the original options. PS. Having internal mobility within the company was helpful too. If someone felt so strongly that they couldn't get onboard with the team consensus, they could go work on a different problem :-) https://lnkd.in/gkjun3KE

8

Interesting conundrum: A kebab menu is more visible using white dots on the blue background than black dots on the same blue background. But the contrast ratio is bad with the white (2.76:1) and good with the black (7.6:1). I’ve asked three separate folks with low vision to tell me which was better for them. They all say “white dots”. Granted, it’s a very small sample. Still. This leaves my client with an interesting choice: pick the option that is more accessible, but not standards conformant, or stick to the standards but end up with an element that is harder to perceive. What would you do? #Accessibility #WCAG

12

23 Comments

Ezra Klein opens his podcast, with AI Expert, Ethan Mollick, with an all too relatable preamble: “I consistently sort of wander up to the A.I., ask it a question, find myself somewhat impressed or unimpressed at the answer. But it doesn’t stick for me. And I think that failure matters. I think getting good at working with A.I. is going to be an important skill in the next few years.” In the next hour, Klein and Mollick go back and forth with some of the most insightful and practical AI guidance I’ve heard.  For example: Ethan Mollick: "A.I. is built like a tool. It’s software…But because of how it’s built, because of how it’s constructed, it is much more like working with a person than working with a tool. In fact, there’s some early evidence that programmers are the worst people at using A.I. because it doesn’t work like software. It doesn’t do the things you would expect a tool to do. Tools shouldn’t occasionally give you the wrong answer, shouldn’t give you different answers every time, shouldn’t insult you or try to convince you they love you. And I find that teachers, managers, even parents, editors, are often better at using these systems, because they’re used to treating this as a person.” See comments below for some of my other favorite sound bites. I definitely encourage everyone to listen. https://lnkd.in/gjmHCEvq

9

2 Comments

"Can you connect me to an AI agent please?" LLMs' ability to "answer questions" via chat bots really captured the imagination of many, and it's quite exciting seeing how rapidly businesses are deploying the underlying tech, into their work via "AI Agents" -- which can go beyond question answering, and take actions to achieve certain goals. AI Agents, appears to be a likely candidate for the form factor for how we'll experience this tech in our daily lives. I really enjoyed reading this brief snapshot of 101 examples, spread across 6 categories of AI Agents: Employee Agents, Customer Agents, Cognitive Agents, Data Agents, Code Agents and Security Agents. https://lnkd.in/eF9dt9MJ

13

AI/ML workers - not yet ✅ I was skeptical about the idea of: - AI Software Engineer - AI Sales Person - AI Marketing Person - etc. Yes, we would automate big chunks of our work with AI/ML. It is already happening massively across most businesses. If you're not doing it, you should be doing it ASAP. BUT giving end-to-end ownership of whole roles... I don't think that we are there yet. This is the same as robots - We don't start from a multi-purpose humanoid robot. We start from specialized machines that solve specific problems. We would get to the humanoid robots... We would get to the AI/ML workers. But we're not there yet. ... P.S. If you need help with building AI/ML automations... RemoteMore has whole teams with tons of experience in this. We can scope and deliver the whole project for you. DM me if this is relevant for you. 🚀 #remotemore #building #AI #ML #software

49

6 Comments

Awesome to see how our Pinterest Engineering friends take advantage of an almost *700x performance difference* between Honeycomb and their Develocity (fka Gradle) dashboards to find outlier users, old machine types, or use of specific JDKs. 🔥 We love seeing our customers use Honeycomb in all parts of their development process, leveraging our support for custom metadata (monotonically increasing, unbounded build identifiers are a perfect stealth example of high-cardinality data) and 💥speed💥 to deliver better experiences for their users. https://lnkd.in/eGcKeUvY

52

2 Comments

Kudos to Google for taking immediate decisive action against unacceptable disruption, defacement, and violation of their employee code of conduct. Here is the (well written) memo in full. You might want to keep it on hand as a template in case you find your company in this situation. "Googlers, You may have seen reports of protests at some of our offices yesterday. Unfortunately, a number of employees brought the event into our buildings in New York and Sunnyvale. They took over office spaces, defaced our property, and physically impeded the work of other Googlers. Their behavior was unacceptable, extremely disruptive, and made co-workers feel threatened. We placed employees involved under investigation and cut their access to our systems. Those who refused to leave were arrested by law enforcement and removed from our offices. Following investigation, today we terminated the employment of twenty-eight employees found to be involved. We will continue to investigate and take action as needed. Behavior like this has no place in our workplace and we will not tolerate it. It clearly violates multiple policies that all employees must adhere to – including our Code of Conduct and Policy on Harassment, Discrimination, Retaliation, Standards of Conduct, and Workplace Concerns. We are a place of business and every Googler is expected to read our policies and apply them to how they conduct themselves and communicate in our workplace. The overwhelming majority of our employees do the right thing. If you're one of the few who are tempted to think we're going to overlook conduct that violates our policies, think again. The company takes this extremely seriously, and we will continue to apply our longstanding policies to take action against disruptive behavior – up to and including termination. You should expect to hear more from leaders about standards of behavior and discourse in the workplace."

2

Google's new approach: yet to be defined “informed choice” instead of cookie deprecation.  I have a suggestion. We Almost Nailed This Back in '99: Learnings from P3P Reflecting on Google's recent decision to postpone third-party cookie deprecation, I can't help but think: we've been here before. Back in 1999, I was part of a team working on the Platform for Privacy Preferences (P3P) project, a major W3C standards effort aimed at creating machine-readable privacy policies for the web. P3P was ahead of its time. We envisioned a world where browsers could automatically read and evaluate website privacy policies, giving users more control over their data. This idea was notably supported by Internet Explorer, the dominant browser at the time. However, P3P had a fatal flaw: trust. Anyone could claim anything in their P3P policy, and there was no way to verify it. I pushed hard for authenticated compact headers, but digital signature standards weren't ready yet. Without this crucial piece, P3P eventually fell by the wayside. Similar Challenges Being Faced Today Fast forward to today, and we're facing similar challenges. Google's Privacy Sandbox and the broader industry efforts to replace third-party cookies are grappling with the same fundamental issues: How do we balance personalization with privacy? How do we ensure that privacy statements are trustworthy and enforceable? Revisiting P3P concepts could provide valuable insights for our current privacy challenges. Imagine privacy policies that are not just machine-readable but also cryptographically signed and verified by responsible authorities. In this world, browsers can enforce these policies automatically, and users have granular control over their data without reading through pages of legalese. They can rely on recommended preferences from trusted authorities like the AARP or Consumers Union. We now have regulations, blockchain technology, advanced cryptography, and a much more sophisticated understanding of online privacy. But the core principles of P3P – transparency, user control, and machine-readability – are still relevant. What’s Next? What’s Possible? We already have added labels for cookies with the HttpOnly, SameSite, and Partitioned attributes. My suggestion is to make this more useful with a digitally signed “nutrition label” detailing how the cookie will be used. Google's decision gives us a chance to get this right. Let's not waste it by reinventing the wheel. Instead, let's learn from the successes and failures of P3P and other privacy initiatives. Browser-arbitrated privacy controls can work with robust mechanisms for verification and enforcement. It won't be easy, but it's necessary if we want to preserve a vibrant open web. The future of online privacy isn't just about getting rid of cookies – it's about creating a new paradigm for data transparency and control. And who knows? Maybe this time, we'll nail it. #cookiepocalypse #thirdpartycookies #google

27

6 Comments