Tanner Burson

This looks like a great project from my friend Haig Douzdjian. I hope that once he finishes the overview he posts some more details about how he processes the implicit feedback (without giving away all the jewels, of course!). Human-in-the-loop feedback is a critical need for #LLM-based systems, but just like with traditional tabular #ML it's really hard to get enough feedback (aka labelled data). Building models on the side, to watch a conversation with an LLM and evaluate the success of each prompt interaction, seems like a great idea. I have to wonder how to cut through the noisiness of the #data well enough to make it valuable to the model. Hmmmmm!

5

1 Comment

I wish I discovered Mermaid's flowchart language in Notion earlier, this tool is so easy and useful to work with. (https://mermaid.js.org/) For example, a lot of our customers have requested Subscription support in RevOps.io. So we're building a Subscription synchronization feature in RevOps.io that tightly couples Renewal and Expansion updates to a Subscription object. Using Mermaid's flowchart language, I can easily show examples of a Customer Lifecycle and a Subscription object. Anyone else finding this tool to be useful?

6

5 Comments

At Forge, our mission is crystal clear: helping software companies build thriving ecosystems through plugins. Here’s how we do it: ͐➡ Technical Enablement & Infrastructure: Everything you need to support your ecosystem from libraries, SDKs, plugin hosting and tooling ➡ Ecosystem Seeding: Getting your ecosystem off the ground with early-stage momentum. ➡ Security, Compliance & Governance: Peace of mind for you and your users. Companies like Shopify, Salesforce, Miro, Microsoft, Slack, Zoom, and Atlassian have already set the gold standard. They’ve shown why extensible ecosystems are a modern blueprint for success. Our goal? To bring this capability to every software company—whether you’re a startup or scaling fast. Extensibility isn’t just a buzzword. It’s a growth engine: 🪄 It fuels revenue. 🪄 It builds long-term scalability. 🪄 And most importantly, it empowers your customers with solutions they never imagined possible. 🔥 Forge makes that possible. #plugins #productlaunch #startup #saas #technology #api

8

We've been making some cool examples of things that can be built with substrate.run lately. Here's a simple chatbot that can help you cook: https://lnkd.in/eEZm-nRJ If you're working on building a chatbot, take a look and see what it might look like to use Substrate 🧑‍🍳

19

2 Comments

Closet Vikings and (decidedly not) Closet JavaScripters => Runic Code is in public preview! Check it out at: https://lnkd.in/ge6GN7qa What is Runic Code? It's the most awesome and, probably only(?), Viking-infused JavaScript code quiz app. What can you expect? Strengthen for all those JS battles where being able to divine the purpose of code could mean life or death! Err... maybe not death per se, but you'll still be able to amaze interviewers, colleagues, and the non-Viking-JavaScripting public with your legendary code reading skills. More updates are to come, but criticism both positive and negative will be considered in the Viking thing (parliament-ish gathering). My battles to come will include tackling: - each bit of feedback, - separation of concerns (I currently have my fetch inside App.jsx), - seeing if the bug where correct answers do indeed show up as wrong, as seen in the video, are happening, - adding a license and contributions.md to add more solved Codewars katas, - shuffling the multiple choice choices (hehe, had to double that word for clarity), - adding local storage to keep track of past scores, - and anything else that comes to mind! A thousand thank you's to anyone whose read so far and followed all my nonsense! Please do check out the repo at: https://lnkd.in/ghtg9MP5

22

8 Comments

TL:DR; I’m looking for a partner to lead Terminal7’s dev & community relations. One that groks youtube, reddit & discord and knows how to exit vim. I always wanted to be a one-man-band and as I’m tone deaf, a real band is out of the question.  What’s left for me is to design, develop and market the devtool I need. That was the plan for https://terminal7.dev and it has to change. Today is the 11111111₂ day since the black sabbath and it's time to acknowledge the war drains too much of my energy.  The sadness is deep and multi-layered: for the kidnapped and their families, for the fallen and for my country. It’s Israel’s darkest hour and our gov is not helping, quite the opposite. Even when one of my posts succeeded in sparking discussion on reddit and selling a few subscriptions, the energy boost lasted less than 8 hours. I was completely depleted when it was published that the hostages Yoram Metzger, Chaim Peri and Amiram Cooper and Nadav Poppelwell are dead. I’ve tried to mute the news feed, but it’s not practical nor moral. As an Israeli I feel obligated to read the stories of the fallen and listen to their grieving families.  Whatsmore, as a father to a 19 year old daughter-soldier I don’t need the news feed to remember the young soldiers held by Hamas: Agam, Daniellla, Karina, Liri & Na’ama. It feels like someone installed an engine of sorrow in my stomach. Sometimes it revs faster and sometimes slower, but it’s always there, draining my energy and happiness. I have little control over this engine of sorrow so I’m looking for a partner to help me keep the revs down.  A great communicator to lead Terminal7's marketing while focusing on community relations. The ideal candidate would have experience building and engaging communities, especially on platforms like YouTube, Reddit, and Discord. The position is flexible - it can be full or part time, hybrid or remote. You can be highly experienced or less so. In any case, the base salary will be modest with significant bonuses tied directly to revenue growth. If that sound right for you, please email your resume to [email protected]

11

Building an MVP shouldn’t take months or millions of dollars. As “The Business CTO,” I’ve built 15 MVPs in record time—often in less than 2 weeks. I’ve even developed a working version of a software in 1.5 months that someone spent 15 years and 19 companies trying to create. I’m getting a lot of DMs from entrepreneurs who want to build faster without sacrificing quality. So, I’ve opened up 5 exclusive 1:1 mentorship slots on Intro. What can we cover together? ✔️ Rapid MVP creation ✔️ Scaling your business with automation ✔️ How to sell tech without getting your development team angry at you ✔️ Product validation and scope optimization Disclaimer – only sign up if you’re serious about building fast. Comment “interested” and secure your spot now: intro.co/jeffzigman

7

3 Comments

I recently read Tilde's blog post, it's a great deep dive into the cutting edge of sparse autoencoder research! They delve into the balance between interpretability and reconstruction accuracy, addressing issues like the "infinite width codebook" and "feature oversplitting" that often hinder robust feature learning. They also explore the Top-𝑘 activation function through the lens of Information Bottleneck theory, showing how it can improve robustness under noisy conditions. If you're interested in mechanistic interpretability or the theoretical foundations of machine learning models, I highly recommend checking it out! https://lnkd.in/eupt7JTs

12

⚡️https://Docsify-This.net v1.9.5 is now available.⚡️ Highlights include: ✅ Added Docsify RunKit plugin (with thanks and appreciation to John Hildenbiddle for developing the plugin) to support embedded interactive #JavaScript REPL (Read, Evaluate, Print, and Loop) blocks😎 ✅ Support for .markdown file extension in addition to .md ✅ Various improvements to the project ReadMe ✅ Minor bugfixes Docsify-This remains the easiest way out there to turn an online Markdown file into a Web page, just follow these steps: 1️⃣ Create or view a Markdown file on GitHub, Codeberg or a Webserver and copy that URL 2️⃣ Go to http://Docsify-This.net, paste the URL into the Web Page Builder, and tap "Publish as a Standalone Web Page" 3️⃣ That's it, no step 3! Give it a try at https://docsify-this.net #Markdown #OER

2

In today's episode of The Daily CTO, I had the pleasure of sitting down with Zachariah Peterson, founder of Northwest Engineering Solutions, to dive into a topic many software CTOs face but few feel fully prepared for: the leap into hardware development. In this conversation, we explore a crucial mindset shift and 5 essential lessons to navigate the hardware space effectively, including: 1. Understanding the real costs and timelines of hardware prototyping 2. Navigating complex compliance requirements for physical products 3. Selecting reliable components that won’t disrupt production 4. Integrating hardware and software seamlessly to create cohesive products 5. Managing production timelines and building scalable, compliant solutions With his deep expertise in hardware consultancy, Zach sheds light on why hardware requires a more deliberate approach than the agile, iterative nature of software. #thedailycto #software #hardware

11

2 Comments

If you're attending Trustmarque Fusion, come and meet Sumo Logic at this one-of-a-kindI come and meet Sumo Logic at this one of a kind tech and leadership conference and learn how the #poweroflogs can fuel #DevSecOps. 🔥#TMfusion24

1

It was hard, but we now have a primitive at Gadget that lets us do absolutely 0 downtime atomic changes to an application's structure or data under the hood, with 0 requests dropped. I call it "request catching" but it was inspired by this Braintree eng post I read long ago For stateless workloads like serverless functions, 0-downtime deploys are pretty easy. You stand up the new one next to the old one, use a load balancer to start sending requests to the new one, and slowly stop sending requests to the old one. But for stateful workloads, it's much harder. You can't just start a new Postgres and send some queries to the new one and some to the old one without lying to somebody about state. If you want to be consistent, there has to be one source of truth. Fancy DBs have quorums for this, but PG doesn't. If you want to be consistent, you're effectively forced to have some tiny-but-real window of downtime in between draining the old one and starting the new one. That said, there's nothing forcing you to return errors to clients while that is happening. If your infrastructure supports it, you can pause requests while this changeover is happening under the hood, and then resume them once it is over. Clients will notice a longer request that normal, but it will be served like normal and they'll be none the wiser! What's required to do this though is a deep deep integration between the load balancing infrastructure and the thing doing the database deploy under the hood. We've built this up using a distributed locking tool, where all requests start by taking out a shared lock. When we're ready to do the maintenance under the hood, the maintenance workflow takes out an exclusive lock, blocking all requests from proceeding. The workflow does the atomic change it needs to make knowing it has exclusive access which means nothing will change under its feet, and then releases the exclusive lock as fast as it can so those requests can proceed. As an optimization, we only check the lock on the request path if an app is marked as near a maintenance window. This is only possible because of Gadget's architecture where we bundle together the load balancer, the request processing layer, and the database access tier such that we can reliably intercept all requests and all database transactions and do this lock checking. Its these nasty, in-the-weeds details that make me so proud of Gadget -- we can do world-class engineering stuff under the hood on behalf of all our apps without them ever noticing. If you're interested in doing this on your stack some OSS tools have the primitives you need! We use nginx for our load balancer, Temporal for the reliable workflow execution and we're currently using postgres advisory locks for hte locking. I think the Braintree post has been lost to time after they got acquired, but @simonw has a great post on the subject here: https://lnkd.in/eZmiktGT

51

I've spent much of the past decade delivering AR platforms, each with diferetn requirements and hardware performance constraints. Lately much of my focus as been getting the best performance and experiences on VisionPro. I will share some tips on performance as well as managing projects as they get larger. First up, consider separating your 2D and 3D views. Often the knowledge, developer skills, and performance ramifications differ widely. Keeping them separated will speed development and reduce bugs.

26

4 Comments

We call this self integration - Coming soon. Tyne integrations can utilize any API documentation, read all functions, create and execute code within minutes instead of weeks which reduces manual labor significantly. For example, we created 80 functions from the Spotify API on single thread in just 10 minutes, a process that typically takes weeks. With multi-threading, this could be even faster. #innovation #digitalmarketing #technology #creativity #Future #futurism #Entrepreneurship #Markets #Startups #Marketing #SocialMedia #VentureCapital #SocialNetworking #LeanStartups #Economy #Economics #sustainability #healthcare #education #design #sales #supplychain #retail #data #top #recent #best #LinkedIn #AI #Automation #Intelligence #Analytics #Artificial #Machine #startup

3

"You can't manage you can't measure." However, the age old question of "How to measure software development productivity" has always been fickle. In the latest episode of Level Up, I had the chance to sit down with Eli Daniel, Head of Engineering at Jellyfish. We talked about why metrics are important for engineering leaders and how to select the right metrics for your organization. Spoiler: DORA isn't a one size fits all. Link to the podcast in the comments, or check it out below 👇

14

4 Comments

Interested in how to kickstart and scale design engineering in your organization? Check out the latest episode with Steve Saul!

9

Had so much fun recording this episode of Jamstack Radio where we got to jam out about local-first software, SolidJS, and other components of our real-time collaborative stack. Brian Douglas is a long-time mentor and friend and one of the most knowledgeable people about new trends and technologies in the industry. https://lnkd.in/etCYRBhG P.S. if you want to get started with open source or get insights into open source projects, you have to check out Brian's company, https://opensauced.pizza. There's nothing that compares.

16

2 Comments

On April 18th, just last week, Meta unveiled Llama 3, lauded as "the most capable openly available LLM to date". This remarkable achievement by the Meta team comes hot on the heels of Llama 2's release last July. Notably, the new models surpass their predecessors and competing offerings from other providers in terms of performance. Meta's commitment to open sourcing these powerful models while prioritizing model safety and responsible usage is commendable. (https://lnkd.in/ghv3RJRz) Yesterday, Microsoft introduced Phi-3 Mini, a member of the Phi-3 family of models, promoted as "the most capable and cost-effective small language models available". According to Microsoft, the key to achieve high performance in such a small package lies in the quality of the training data. The model has been open sourced and is now available on Ollama and Hugging Face. (https://lnkd.in/gKgSmrBn)   These compact yet powerful models serve as invaluable tools for creators eager to bring their visions to life or simply experiment with cutting-edge technology, particularly for those with limited access to resources.   Shortly after the release of Llama 3, reports began surfacing of enthusiasts running the 8B model on a Raspberry Pi 5 equipped with just 8GB of RAM—a modest $80 single-board computer. (https://lnkd.in/gbeQPhQ2) While the performance understandably reflects the limitations of such a platform, the fact that the model runs and generates an output is nothing short of magical. Similarly, the Phi-3 Mini model can run on ubiquitous devices such as smartphones.   Reflecting on my own journey, I recall the pivotal role that early access to a low-cost computer played in shaping my passion for coding. The gift of an 8-bit Commodore 64 home computer ignited my curiosity at a young age. Later, with the guidance of an outstanding Computer Science teaching staff at my high school IISS Marconi-Hack Bari, I embarked on ambitious projects like a networked multiplayer version of Battleship written in Turbo Pascal and 8086 Assembly, complete with custom sprites and lo-res graphics. These formative experiences paved the way for my career, eventually leading me to various engineering and leadership roles.   It all began with a simple, low-cost, modestly capable computer used for experimentation—a gateway to discovering my life's passion.   I encourage educators and parents alike to consider setting up similar low-cost experimentation environments. By introducing younger generations to open-source AI technologies, we can nurture their creativity and help them uncover their true passions. It's not just a weekend project—it's an investment in their future and the future of innovation and discovery. #education #youth #AI #LLM #future #innovation

24