flâneur

Selene Xu

10 followers · 13 following · 397 views

on the atlas — 39

highlights — 46

  • Big ambitions, low expectations, and high standards are a powerful combination for living your best life. Ambitions pull you forward when it's hard. They connect you to something larger. One of my most important ambitions is to be a great father and friend. Another is to leave the world a better place than I found it. You can't have a meaningful life without a connection to something larger than yourself.
    Post | Feed | LinkedIn
  • This is fundamentally what startups are all about. Constantly ploughing through and getting pummeled and rinsing and repeating. And this can be depressing for some founders. And, this is not for everyone — and that’s ok to move on. But if you’re still excited to try to surf your startup wave, keep going! Everyone gets pummeled over and over again — you’re not alone — but just get back up on the board and continue. You’ll figure it out.
    I’m (still) not crushing it – Elizabeth Yin
  • This is fundamentally what startups are all about. Constantly ploughing through and getting pummeled and rinsing and repeating. And this can be depressing for some founders. And, this is not for everyone — and that’s ok to move on.
    I’m (still) not crushing it – Elizabeth Yin
  • No phone before bed. I keep my phone in the kitchen Use phone meetings to walk / jog to get exercise – there’s often no other time Reduce meetings with voice messages via Async.com Sleep early and wake up early to get a better handle on the day
    I’m (still) not crushing it – Elizabeth Yin
  • People have an enormous capacity to make things happen. A combination of self-doubt, giving up too early, and not pushing hard enough prevents most people from ever reaching anywhere near their potential.
    How To Be Successful
  • The eminent, on the other hand, are weighed down by their eminence. Eminence is like a suit: it impresses the wrong people, and it constrains the wearer.
    The Power of the Marginal
  • on how the consumers of LLMs — or how the users of LLMs — will actually interact with those LLMs. And more specifically, who will own fine-tuning.
    Sarah Catanzaro — Remembering the Lessons of the Last AI Renaissance | gradient-dissent – Weights & Biases
  • And so I definitely saw eye to eye with you guys on that. I also think that if you ask people today and I ask the portfolio companies that we get to work with, to talk about their machine learning strategy, not always, but very often it ends up being a second step in their company strategy, which is, oh, when we can afford a team of a bunch of data scientists and we can really resource this, then we’ll invest in it.
    Self-Serve Apps for ML Teams | Greylock
  • upstarts have found opportunities to pick off and go after specific pieces of the payvidor services map
    Payvidors, Unbundled: Opportunities in Healthcare Fintech | Andreessen Horowitz
  • The unique laws of physics of our $4 trillion healthcare system, primarily a result of third-party payor (e.g. insurance carriers, self-funded employers, and government entities) and esoteric regulatory dynamics, make healthcare a hard market for a generalist company to go after.
    Payvidors, Unbundled: Opportunities in Healthcare Fintech | Andreessen Horowitz
  • the primary battle between every startup and incumbent is whether the startup gets distribution before the incumbent gets innovation
    Payvidors, Unbundled: Opportunities in Healthcare Fintech | Andreessen Horowitz
  • Kiser explained to FreightWaves that while many larger firms are beginning to use application programming interfaces to exchange data, he recognized a number of supply chain participants would continue to use EDI to communicate with one another.
    Orderful raises $19M to simplify EDI integrations for logistics community - FreightWaves
  • While the front end of commerce has seen significant innovation over the past decade, much of the back end of the supply chain is built on disparate systems and manual processes. We see significant investment opportunities in the modernization of the supply chains, and Orderful provides a leap forward with EDI integration
    Orderful raises $19M to simplify EDI integrations for logistics community - FreightWaves
  • . Today, fine tuning on domain-specific data is one of the key ways to properly “steer” the model and stop it from generating gibberish. The extent to which fine tuning is needed in the future will closely parallel the rate of improvement in underlying models like Stable Diffusion and GPT-3. We’re watching this pattern closely as it will determine the extent to which the locus of value moves from the vertical to the horizontal, and whether fine tuning is a credible route to defensibility in the generative AI space.
    Foundation Models Are The New Public Cloud | Scale Venture Partners
  • Generation, not just Classification
    The Age of Open Foundation Models
  • . Generative systems — ones that automatically produce text and images from simple text prompts — have advanced to a level where they could have wide-ranging business uses. A
    AI’s sudden big leap forward into usefulness | Financial Times
  • Enterprise ROI. We have seen LLMs be leveraged for tasks like graphic design or for more complex tasks like incident response in DevOps (like Zebrium). The latter category of business is long-term more exciting – though a much harder model and product to build in the short-term – because the value of the work is much higher ROI and therefore willingness to pay from enterprises will be much higher. Humans in the loop. Building on the above point, we love companies that give knowledge workers leverage on their time and allow them to make better decisions or produce better content. Additionally, …
    The Promise and Perils of Large Language Models | Two Sigma Ventures
  • At BCV, we view large language models as a similar foundation with broad implications, and we’re particularly excited for the implications in B2B software.
    Large Language Models Will Redefine B2B Software - Bain Capital Ventures
  • Cambrian explosion of startups building applications on top of LLMs including everything from copywriting platforms to developer tools.
    Large Language Models Will Redefine B2B Software - Bain Capital Ventures
  • My perspective is that open-source models will come to dominate the LLM space as opposed to proprietary models. Earlier this summer, BigScience, a LLM research workshop, released the BLOOM model, which was the first open-source, multi-lingual large language model, trained on $7M of publicly-funded hardware cost. This came about two years after the initial release of OpenAI’s proprietary GPT-3, which has similar capabilities. The one-time nature of training costs will continue driving the open release of future models for more broad and cost-effective public benefit.
    Large Language Models Will Redefine B2B Software - Bain Capital Ventures
  • Separately, for the LLM applications that own the models themselves, will training & inference costs be feasible to continue paying (even if they continue to decrease)
    Overview & Applications of Large Language Models (LLMs)
  • will it all be commoditized by many providers offering similar models, or will the most cutting-edge company (with the best engineers, hardware, data, compute, & community) become a gatekeeper?
    Overview & Applications of Large Language Models (LLMs)
  • Product insights Viable, Enterpret, Cohere, & Anecdote organize & summarize product feedback from users (e.g. support tickets, surveys, analytics) into actionable insights for future product development.
    Overview & Applications of Large Language Models (LLMs)
  • pain points include wanting to self-host or fine-tune their own models, customize workflows, as well as wishing to fix some challenges Codex has with frontend frameworks & test generation.
    Overview & Applications of Large Language Models (LLMs)
  • How mission critical is the LLM for the business?
    Overview & Applications of Large Language Models (LLMs)
  • It’s important to note that the competitive advantage isn’t just the private data used to train the model initially, but the additional data you get when customers interact with the model, telling it what is right, wrong, and sometimes what the answer should be instead.
    Overview & Applications of Large Language Models (LLMs)
  • I constantly have to remind myself that “modern” ML is nascent - 2011 was the first year a convolutional neural net (a type of deep learning model) won the most popular computer vision competition. Transformers, which power the LLMs mentioned above, were introduced by Google Brain in 2017 - just a few years ago!
    Overview & Applications of Large Language Models (LLMs)
  • Importantly, each of these cycles took years for the tech to mature enough for many successful companies to be built. Given the rapid pace of innovation in machine learning (ML), especially in LLMs, I believe that we are beginning a wave where many significant companies supplying or utilizing this technology will be built.
    Overview & Applications of Large Language Models (LLMs)
  • So we can say all of the sort of tried things at this point like I do believe that natural language is this new interface that's real and that's going to happen and you know we will control lots of things not everything because sometimes it is better but we will control lots of things by like telling a computer what we want.
    Live Fireside: Sam Altman, CEO of OpenAI, w/ Elad Gil
  • o we can say all of the sort of tried things at this point like I do believe that natural language is this new interface that's real and that's going to happen and you know we will control lots of things not everything because sometimes it is better but we will control lots of things by like telling a computer what we want.
    Live Fireside: Sam Altman, CEO of OpenAI, w/ Elad Gil
  • What do you think are going to be the characteristics of the application level companies for AI that are going to work like what are the common threads that people should be looking at or looking for.
    Live Fireside: Sam Altman, CEO of OpenAI, w/ Elad Gil
  • I think we haven't seen this sort of like application level companies emerge yet or at least not like I think there will be new trillion dollar AI application companies that start my guess is they haven't quite started yet. Maybe the first ones out but I would bet in the next few years we see them get started. And I think we haven't been able to say that any category for a while. Like we are far from the first people to observe that the 2010s where this like weird mix of investors making ton of money because we're in this like up elevator evaluations when nuts.
    Live Fireside: Sam Altman, CEO of OpenAI, w/ Elad Gil
  • eah I mean I think OpenAI is this weird like kind of one thing so I wouldn't try to like draw too many conclusions from it. I think we haven't seen this sort of like application level companies emerge yet or at least not like I think there will be new trillion dollar AI application companies that start my guess is they haven't quite started yet. Maybe the first ones out but I would bet in the next few years we see them get started. And I think we haven't been able to say that any category for a while. Like we are far from the first people to observe that the 2010s where this like weird mix of …
    Live Fireside: Sam Altman, CEO of OpenAI, w/ Elad Gil
  • Yeah I mean I think OpenAI is this weird like kind of one thing so I wouldn't try to like draw too many conclusions from it. I think we haven't seen this sort of like application level companies emerge yet or at least not like I think there will be new trillion dollar AI application companies that start my guess is they haven't quite started yet. Maybe the first ones out but I would bet in the next few years we see them get started. And I think we haven't been able to say that any category for a while.
    Live Fireside: Sam Altman, CEO of OpenAI, w/ Elad Gil
  • business value transition and these unbelievable research advances, these impactful generative applications have been limited mostly to large internet companies and a few well-funded independent lab
    The Age of Open Foundation Models
  • . Prediction and generation are core parts of intelligence, and these tasks are a significant expansion of the AI field's scope: for example language generation, code generation, voice generation, image generation, and even math solving.
    The Age of Open Foundation Models
  • e pre-trained, transformer-shaped "foundation models" that will generalize well to many tasks either out of the box or tuned with dramatically fewer/more efficient training samples
    The Age of Open Foundation Models
  • Many of the core ideas in ML, such as stochastic gradient descent and backpropogation, existed in obscurity for decades.
    The Age of Open Foundation Models
  • And yet...until now there have been few tinkerers, people testing the bounds of what products you can build, of startups and side projects and cool demos. Why? It's been an exciting week (really, decade) in ML and I have a hunch the game is finally changing.
    The Age of Open Foundation Models
  • We are entering the age of assisted knowledge work and play. After I first wrote about the conversational economy and the potential of AI agents, there was a 5Y+ period of disillusionment. Bots in the wild, for example to replace human support agents, didn’t live up to expectations. But we have seen a flurry of applications of AI that have had counterintuitive success. Instead of fully automating repetitive work, they’ve excelled at higher level, counterintuitively creative, nuanced and empathetic tasks as varied as art generation, language translation, and personalized mental health coaching …
    thinking with machines
  • Fascinating. So as we get close to the end of allocated time here, a couple of questions from the group. Rachel asks, “Which is the better path towards online learning? Adapting infrastructure to train the DNN in an online way? Or different traditional algorithms like bandits?” I think it’s a really interesting question. I think it really depends on the use case and on data. I think a lot of companies that I see… When they do online learning, they try to start with a simple use case, but I’ve seen people who try to start with a similar use case and still run into the same problem as people who…
    In conversation with Chip Huyen, Writer and Computer Scientist – Matt Turck
  • Sounds good. One more. What are some applications that are leveraging online learning today, e.g., recommendation systems for e-commerce, and how is this implemented on device or cloud?
    In conversation with Chip Huyen, Writer and Computer Scientist – Matt Turck
  • And Jack Hanlon in the comments, says, “Chip is so on point here. Online learning is much harder. Can confirm as we’re getting into more and more of it at Reddit.”
    In conversation with Chip Huyen, Writer and Computer Scientist – Matt Turck
  • Yeah, so I was talking about online predictions. So how is this hard? So to do online predictions, you actually need two components. So the first component is you want a model to make predictions very fast, like fast inference, like low-latency. So first of all, you don’t want to open Netflix and the webpage load for a minute and to show you recommendations. So maybe for a certain system, it can be very slow and companies don’t want to make users wait. So they generate predictions offline and whenever users have a query, they fetch a query because the time it takes to fetch a query is much, mu…
    In conversation with Chip Huyen, Writer and Computer Scientist – Matt Turck
  • I think it’s very interesting. It is both overdeveloped and underdeveloped. It’s very crowded, but it’s still underdeveloped. My theory is that… and it can be totally wrong and I’d love to hear your thoughts on it. So my theory is that a lot of people try to pluck low-hanging fruit right now. So there are a lot of low-hanging fruit tools, and they are very similar. But there are a lot of big challenges. What’s an example of a low hanging fruit? Data labeling. How many data labeling tools out there? There’s a lot. It’s not a bad thing. It’s just a lot of low-hanging fruit, so there’s so many di…
    In conversation with Chip Huyen, Writer and Computer Scientist – Matt Turck
  • In theory, an artificial general intelligence could carry out any task a human could, and likely many that a human couldn't. At the very least, an AGI would be able to combine human-like, flexible thinking and reasoning with computational advantages, such as near-instant recall and split-second number crunching.
    What is artificial general intelligence? | ZDNET