Justin Meier

Hello, I'm Anthony and I'm looking for a A+ killer Head of Growth for Equip Foods! Here are some wild stats on Equip: 🚀 Mid 8-figures in revenue, bootstrapped, and highly profitable (no VC BS) ✌️Only 2 FTE at the moment (no micromanaging or endless meetings) ❗️Only currently focusing on one marketing channel (affiliate/influencer) 🛍️ Industry best CR, LTV, and subscription #s (clear product market fit) 🥤Real food supplement products that are clean, transparent, and of the highest-quality that people absolutely love (actually doing something important in CPG) Perfect role for anyone who gives a damn about health, wants to get their hands dirty and build out a growth process, have huge amounts of agency and autonomy, and keen to make a dent every day when they show up to work. Here's a link to why we exist: https://lnkd.in/gGb-qCA4 Here's a link to the job description and app: https://lnkd.in/gxwE7xk7 (TIP: DO NOT APPLY IF YOU ARE PLANT-BASED) Please tag anyone you think could be a good fit! (P.S. I've been a tad quiet here. Are you interested in anything else about this business? Should I share more on LinkedIn?)

74

14 Comments

Testing LLM applications is hard, but robust regression tests are critical for developing software. I wrote a short blog talking about how we are doing this at Eidolon. Give it a read if you are having issues here! And please let me know what y'all have been doing so we can improve here as well.

5

1 Comment

TL;DR: I'm still #OpenToWork for my next chapter where I can lead product strategy for highly impactful products (ideally in bio-, med-, or health-tech)! If this sounds like something you know of, please reach out (and save me from the insanity that the internet has become) 🤓 Continue reading for today's comedy of errors... #DITL of the job search in 2024: Found a job that looks interesting. Pay is about right. Started to apply - oh good! LinkedIn integration should make it quick. Oops, I need an account. My email is already in use? Hmm, I don't remember applying to this company before. I'll reset my password. Of course my randomly generated password doesn't meet the (arbitrary) criteria. Finally, no errors. Just kidding! You can't change your password more than once in 24 hours for ~security~. Well, onto the next… (As an aside - forever thankful to my password manager for keeping all my passwords straight for the eight different companies all using #workday for their applicant portal...)

52

3 Comments

I'm finally ready to share an experiment I've been working on with the Future Audiences team at Wikimedia. It's a pretty special team where I get to have space to think about the intersection of things like LLMs and the evolving online space of (mis/dis)information: specifically how can this technology not replace humans as an information source, but help better connect us to living sources of information created by humans. I have this hope/dream that tools for us to verify information in a world where creating questionable information is so widespread will be as easy and normal as your internet browser coming with built-in adblockers to protect you from being overwhelmed with pop-up window spam. Anyway, the experiment we have is called Citation Needed, and if you are using the Chrome browser, I'd love for you to try it out. After you install it, you can use it to check short texts online to see if they are corroborated by information on Wikipedia, and have the opportunity to learn more about it. It's still experimental, and uses LLMs/AI, so you might see some funny behavior once in a while -- which you can report to us in the extension -- though I invite you to imagine what this functionality could look like in the near future, and/or beyond the confines of a browser extension. And then let us know where this sci fi daydream takes you. Link: https://lnkd.in/eQwd4iqK

96

8 Comments

If you missed Google I/O '24 then watch this recap to catch up on the latest news, announcements, and AI updates from Google:

What gets repeatedly lost in this conversation is the difference between platform companies and product companies. Platform companies like OpenAI/Microsoft/Google/Anthropic, will capture extraordinary value from GenAI because the cost to play is so high (creating barriers to entry) and the use cases are varied. It���s no different than the desktop revolution, the mobile revolution, and the cloud revolution. Yes, for those small number of players the potential value creation is nearly limitless But for the rest of the companies in the world, who build products that serve a relatively targeted set of users and problems, any reasonable person would say “where’s the beef?” How has GenAI actually changed a manufacturing business or a travel company? If you’re being honest, it hasn’t. Maybe it will, but I haven’t seen evidence yet. The irony of course is that the original poster here left a platform company (Microsoft) for a product company (Qualtrics).

4

1 Comment

How can we speed up the training of large language models (LLMs)? This issue comes up when people want to use LLMs for specific use cases. I attended a meetup on that included this very topic. Typically you have to add people who know what the right answers are into the loop. It turns out using another LLM to do the review can speed this up. The new process is called Reinforced Learning AI Feedback (RLAIF). I linked to a paper on this topic in the top comment.

3

1 Comment

Ever been curious how different LLMs stack up on coding and code understanding tasks? Want to experience for yourself why context is so important to getting high quality answers? Check out https://s0.dev! Our awesome DevRel team at Sourcegraph have created the "LLM Litmus Test" which allows you to directly compare the top LLMs and see which is better for your use case. You can even toggle on/off Cody context to get a feel for how answers change! Big shout out to Chris Sev and Ado Kukic for the work!

30

2 Comments

10 months ago, I embarked on this journey with a simple yet ambitious goal: to create an AI assistant that knows you like your best friend and can truly enhance your life. Now, Ario is in open beta! 🎉 This journey has been filled with - Joy from hearing beta users say, "Honestly, I'm stunned, like wow," - Excitement from rounds of rapid iterations - Sudden enlightenment in solving unexpected challenges while building our AI agent Every step has brought us closer to creating something truly special - Ario's Personalized, Proactive and Private. I’d love for you to give Ario a try and let me know your thoughts. Your feedback helps us shape Ario. And if you're also navigating the fascinating world of AI, let’s connect and exchange insights!

20

1 Comment

Launch day is finally here! You can reserve Quilt starting today, and installations begin later this summer in the Bay Area. I am so proud of this team and the product we have all been building with care over the last two years. I’m blown away by what dedicated, talented people can accomplish when working together on something that matters to them. Quilt is the smartest way to heat and cool your home. The product is beautiful, smart and efficient - better than anything else on the market. We’ve also made it the most straightforward system to buy. A little backstory… Purchasing a heat pump can be pretty daunting. I know because part of what started me on this journey was struggling to buy one for my home a few years ago. I couldn’t believe how hard it was just to get basic information. It took me weeks of having contractors visit, only to be told completely contradictory information (heat pumps don’t work, or they do work but they won’t work with my ductwork, or they will work with my ductwork but I need to widen it, I need three units, or one…on and on). The pricing I got was utterly all over the map with apples-to-oranges quotes. I was often never shown any information about the actual product I was getting. And information about rebates was confusing and I almost missed out on some. That’s why we’ve launched Quilt the way we have: with clear, transparent pricing that includes the cost of installation. We surface all the incentives and rebates available to you based on your location and the circumstances of your home, and discount them upfront off the cost. It feels more like how you expect to buy things in the age of the internet. It’s what I wished had existed when I was looking for a heat pump. There’s such a bright future ahead as we transition to a clean energy economy. A big part of my optimism around this comes from the opportunity we have to make things not just more sustainable, but simply better. Quilt is a super smart (and seriously beautiful) way to get your home off of fossil fuels and up your home comfort level at the same time. It’s an upgrade in every way. I’m grateful to the whole team who made this happen, and in particular want to thank Andrew Carl, Claire Gagne and David Bedingfield for their incredible work on our reservation experience. I’m also so lucky to work with my two inspiring co-founders, Paul Lambert and Matthew Knoll. What a journey. And it’s just getting started. Visit Quilt.com to learn more and reserve yours today. The future of home heating and cooling is here.

127

12 Comments

Excited to represent FosterHealth AI at American Society of Clinical Oncology (ASCO). Curious to know how you can use AI in your practice? Let’s connect to chat more on AI-powered medical documentation and research! #ASCO2024 #HealthAI #AIforResearch #AIforDocumentation 

31

I wrote a bit of a gripe about no-code stuff last week and people have flagged the future of AI software engineers to me. Yes, we will get to the point where a whole ton of the "code writing" is done by AI... and the sooner the better from me - anyone who's written enough code for the novelty to have worn off would very much like to write less of it. But the writing part isn't the hard part. The hard parts are figuring out what you want computers to do, how you want them to interact with the world, in what order, with what constraints, etc. Once you've done that writing the code is just how you tell computers what you want. And whether you tell them in Python or Java or R (if you're mental) or plain English through an AI doesn't matter, you're still coding.

11

2 Comments

Let me be clear that I'm not a gaming expert, *but* having written some bots for empathetic interviewing, I've got a feel for how gAI fits into interactive gaming. Also have a feel for how temperature settings can influence the experience: too cold, and you're talking to a canned robot; too hot and it's loaded with ADHD non sequiturs. This said, I'd suggesting mixing it up: the real world is a mix of hot-heads and cold fish, so use some of each. If you have speciality roles (say a real trickster, or a genuine confidante), then find people you want to emulate to help you build training sets, then do lots of alpha testing to get the impressions of real humans. Final thought is that I've begun messing a bit with meta-adaptive behavior. If you chat with a stranger on the street, there are a few phrases that might get that stranger to warm up to you, and other phrases that freeze things up. The capacity to adjust temperature (or even core model) on the fly could make for a fascinating virtual experience...

6

2 Comments

Two truths and a lie 🔮: 1) Our team is the perfect mix of every synonym of passionate, cerebral, and customer-led. 2) We’ve raised $25M from some of the best at General Catalyst and Forerunner. 3) Fay is not hiring and you shouldn’t consider applying. I love our team. Yet, still, the whole is greater than the sum of its parts. Citing the team as the reason a startup is doing well often sounds like a cop out answer (even if it’s true), but what isn’t talked about enough is how cool it is working with a team of people all hungrily chasing the same thing you’re chasing. For everyone at Fay, working here is a blessing. We’re blessed with the opportunity to sprint the marathon of building a company solving the food and nutrition crisis in America. ❤️ We’re not just allergic to average; we set the bar high and keep raising it. If you’re hungry to accelerate your career and want to improve & save lives while doing it, I entreat you: apply to Fay. Who I want sliding into my DMs 🫶: - Operations - Growth - Head of Finance - Payor Strategy - Provider Sales - Enterprise Sales - Corporate Development - Chief of Staff - Software Engineer - First iOS Engineer - Product Manager Come find out why we’re one of the fastest growing startups around and the best team to ever do it. I can’t wait for you to meet everyone. https://lnkd.in/eGb9qZj3

52

11 Comments

🚀 Exciting News! 🚀 I'm thrilled to announce the launch of a groundbreaking Authority Status feature for the Sunshine Law Library of the Kentucky Open Government Coalition! Leveraging the power of GPT-4 Turbo, we’ve extracted citation treatments and generated explanations for every Attorney General Opinion and Decision published since 1977. Now, our users can easily see when an older record has been modified, overruled, or withdrawn by a later authority. To our knowledge, this is the first time such a feature has been offered for Kentucky Attorney General records. Traditionally referred to as "Shepardizing" or "KeyCite" by major legal publishers, this valuable tool is typically restricted to court records and available only through costly subscriptions. We’re incredibly proud to offer this service for free, supporting our mission to promote and preserve Kentucky’s open records laws. Learn more about this exciting launch in our blog post: https://lnkd.in/euT7VEAN #OpenGovernment #LegalTech #AI #KentuckyLaw #PublicService #Innovation

7

1 Comment

Some context from the #Founder perspective. As liberal as Seattle is as a city with policy and politics, it's incredibly conservative from an investment standpoint. One of the reasons you may not hear big "home-run" type pitches or pitches with sizzle, is simply because investors in Seattle don't like those types of pitches and Founders have been ingrained to not present in that manner. They may say they do, but ask any Founder and they'll tell you how they have been told to size down their pitch or talk about MVPs, first customer, first revenue, etc. or how what matters more is what you can do first couple years rather than the home run. Those early metrics are important by the way, but different from the big home run. Do you know who is receptive to homerun pitches? Investors in the Bay, NYC, Miami, Hong Kong. Seattle is not great for early-stage Founders with many prioritizing raising outside the city. We will be raising capital in the fall and as of now, have no plans of reaching out to Seattle-based VCs. I should also add that our company has a blockchain component and sadly Seattle VCs have not embraced crypto or blockchain. Welcome to #Seattle, Kyle Lui! Hopefully, you can help change the narrative! GeekWire

21

7 Comments

When your favorite author is turned to a robot, what do you do? You contribute to its codebase. A bit over a decade ago, I read that the cognitive science dept at University of Memphis was building an android version of one of my favorite authors, Philip K Dick. He's a prolific sci-fi writer, penning works such as "Do Androids Dream of Electric Sheep?" and "A Scanner Darkly". Even though he lived over 75 years ago, his works tackle the disintegrating boundaries between human and artificial intelligence. When I heard about the grant, I emailed one of the professors and spent the summer in Memphis, writing XML code for the android's chat function. Unlike LLMs, which digest enormous volumes of data and use probability to generate an output, PKD used strict rules to respond to users: parsing the grammar of the user's input, and massaging it into a response. Q: How are you today? A: I'm happy it's sunny. Q: Why are you happy it's sunny? A: I get to go outside. Q: Why goes going outside make you happy? etc. This was a form of "symbolic" AI (rules and systems based). For the first 50 years of AI's development as a field, it was studied much more rigorously than "stochastic" AI (data + probability based). It spurred major hype cycles when researchers were able to create machines to beat chess grandmasters and plan travel itineraries, but fell by the wayside when people saw the limits of these systems. They could follow rules, but were incapable of "creativity". Today we're witnessing an entirely new hype cycle around stochastic AI . Thanks to advancements in ML and the oceans of digital data we have access to, we can interact with AI that sounds truly human. But users are growing frustrated that these systems can't "reason" or "follow rules". There's plenty of LinkedIn content showing ChatGPT breaking with simple instructional prompts. I'm curious to see if the pendulum will swing back to AI's symbolic origins.

8

2 Comments

Something I re-learn every few years: growth happens when you focus on your core product. 🌱 This time the reminder came from a recent Lenny's Newsletter podcast with Twitter’s former head of product Kayvon Beykpour. Kayvon got Twitter shipping again: Twitter Blue, Live Video, Super Follows, to name a few. But it was the focus on optimization of the core product and its essential building blocks—sign up, onboarding, notifications—that drove growth. That’s the product work you won’t see in a keynote presentation or a Product Hunt release, but that compounds into real, dependable growth. If you’re a PM looking for your next big growth unlock, remember that there’s a good chance it’s right beneath you. The best PMs obsessively study and craft their core product, because they know those micro optimizations add up in a big way in the long run. 📈

36

7 Comments