Brent Zucker

Insightful. Will wait for the My two cents on why it is difficult or going to be difficult - Evolving Complex Application Profiles - Difficulty in keeping it simple - Guardrails being too hard on developers - Accountability (Centralised vs Federated Operating Model) - Inability to measure micro-productivity - Organizations will need custom metrics to measure across concept to design/code to test to deploy to run. (Value Stream Mapping) The cultivation process is still on.

17

This week at Blockfuse Labs, we were introduced to Solidity. We learned about Primitive Data Types, Variables Declaration, Constant Variables, Immutable Variables, Reading & Writing to a State, Ether & Wei, and Gas & Gas Price. We had practical examples and hands-on exercises that helped us understand these fundamental concepts. We wrote and deployed our first smart contracts using Remix IDE, experimented with different data types, and gained a deeper understanding of how to interact with the Ethereum blockchain. Fingers crossed to my blockchain journey that is just beginning 🚀 Blockfuse Labs #web3journey #blockchain #WEB3 #Solidity

8

1 Comment

My colleagues are often involved in modernizing legacy systems, and our approach is to do this in an incremental fashion. Doing this with a mobile application raises some specific challenges. Matthew Foster and John Mikel Amiel Regida share their experiences of a recent engagement to do this, shifting from a monolithic legacy application to one using a modular micro-app architecture. https://lnkd.in/erGxWmy9

1,114

42 Comments

If you are working with Docker... I highly recommend looking at the Slim toolkit. https://lnkd.in/gX_Gb9Kf Why? Slim allows developers to inspect, optimize, and debug their containers using its xray, lint, build, debug, run, images, merge, registry, vulnerability (and other) commands. It simplifies and improves your developer experience building, customizing and using containers. It makes your containers better, smaller and more secure while providing advanced visibility and improved usability working with the original and minified containers. Don't change anything in your container image; minify it by up to 30x to make it secure, too! Optimizing images isn't the only thing it can do, though. It can also help you understand and author better container images. Keep doing what you are doing. No need to change anything. Use the base image you want. Use the package manager you want. Don't worry about hand-optimizing your Dockerfile. You shouldn't have to throw away your tools and your workflow to have small container images. Node.js application images: from ubuntu:14.04 - 432MB => 14MB (minified by 30.85X) from debian:jessie - 406MB => 25.1MB (minified by 16.21X) from node:alpine - 66.7MB => 34.7MB (minified by 1.92X) from node:distroless - 72.7MB => 39.7MB (minified by 1.83X) Go application images: from golang:latest - 700MB => 1.56MB (minified by 448.76X) from ubuntu:14.04 - 531MB => 1.87MB (minified by 284.10X) from golang:alpine - 258MB => 1.56MB (minified by 165.61X) from centos:7 - 615MB => 1.87MB (minified by 329.14X) Rust application images: from rust:1.31 - 2GB => 14MB (minified by 147.16X) Java application images: from ubuntu:14.04 - 743.6 MB => 100.3 MB

4

Last week, I talked to the VP of Engineering at a publicly traded US bank. Our team is consulting them on implementing GraphQL federation, here’s what’s most interesting about the roll-out: Some context around the bank: - 10k+ engineers - 36k API endpoints - 4k microservices - *Substantial* tech budget & annual infrastructure costs Here’s why they want to switch to GraphQL: 1/ They value speed. For them, every request needs to be 10 milliseconds or less. The fact that Grafbase was built performance-first really resonated with their engineering team. 2/ They have a ton of microservices and thousands of REST endpoints. So, they’re struggling not just with discoverability from the consumer perspective, but also with keeping all the endpoints secure. 3/ Because they have so many endpoints, they also have an enormous infrastructure footprint that’s costing them almost $1B per year. Their goal: Save 5-10% infrastructure costs (8-9 figures per year) How realistic is it? I’d say you can even save more money if you go all-in on GraphQL Federation, but only if you commit to this for a long enough period of time. Because here’s what’s going to happen (in their case): In year 1 and 2, infrastructure costs will go up (because they’re adding subgraphs that are talking to their existing microservices). In year 3 or 4, they’ll start to collapse these microservices into *just* subgraphs. And that’s where they’ll see the cost savings. To make this change happen, we are currently helping them work through these two main questions: 1/ How big should the subgraphs be? I told them there’s no absolute number here. But a good rule of thumb is that the traffic shouldn’t surpass 20-30% per subgraph to keep things balanced. Also, similar entities should be grouped together on one subgraph to avoid scattered data across boundaries. If you have credit card information saved on one subgraph, you shouldn’t mix it with other non-sensitive data. 2/ How do we implement a solid authorization strategy? There are tons of authorization strategies (RBAC, ABAC, Zanzibar, …). But the issue for enterprise companies like them is that none of these strategies fully suit their needs. After talking to their engineering team, we built our gateway hooks feature that allows them to write their own authorization strategies using WebAssembly. – There are a lot of moving pieces when rolling out GraphQL at scale, anyone else with experience at the enterprise level here?

53

3 Comments

Excited to announce Lagrange Labs' $13.2M seed round, led by Founders Fund, with support from Archetype, 1kx, Maven 11, Fenbushi Capital, Volt Capital, CMT Digital, Mantle Network Ecosystem Fund, and various angels. Read More: https://lnkd.in/gvxFPH_B Lagrange was founded almost two years ago with the vision of using zero-knowledge proofs to expand the design space of applications that could be built onchain. We saw a simple problem: Developers could not securely use historical or cross-chain state. The name “Lagrange” was inspired by mathematician Joseph-Louis Lagrange, who discovered Lagrange Points, positions of equilibrium in space between two celestial body systems. These points of equilibrium were the perfect analogy for trustless interoperability in crypto. While Lagrange has moved away from being a purely cross-chain focused company, the name remains, as does our goal of bringing equilibrium and stability to onchain applications. The original vision of secure interoperability still lives on within our State Committee protocol. As we’ve progressed, we have crystallized our mission: to unlock new types of data-intensive and cross-chain applications by enabling verifiable computation over blockchain data at a ‘big data’ scale. The way we achieve this is through our hyper-parallel ZK Coprocessor. Over the past few months we’ve hit a number of major milestones as a company: • Launching Lagrange State Committees as the first ZK AVS on @eigenlayer mainnet. Hitting over $5 billion in restaked ETH after only 3 weeks. • The Lagrange State Committees have generated over 30,000 state proofs and have finalized 2,000,000 blocks. • Launching our first ZK Coprocessor testnet, ‘Euclid’. We’ve currently generated over 400,000 proofs at a growing rate of 127,000 proofs per week. • Launching our infrastructure as a decentralized network with 40+ node operators. • Our first paper, Reckle Trees, authored by Charalampos Papamanthou, Shravan Srinivasan, Nicolas Gailly and Andrus S. from our team was accepted into both CCS’2024 and SBC’2023. More to come soon 🤫 None of our progress would be possible if it wasn’t for the team behind Lagrange. A special thank you to the whole Lagrange family whose tireless work brought us to this point. Thank you to all of our friends, partners and investors who have supported Lagrange in getting here. We have an ambitious roadmap ahead of us this year and we look forward to achieving it all of you on our team.

62

9 Comments

I’ve recently had conversations with the Engineering team at a major US Bank about GraphQL Federation authorization strategies. Here’s how it led us to allow teams to customize their GraphQL gateway using WebAssembly: First, some context: We’re currently talking to a bunch of enterprise companies - one of them being a leading financial services company with a team of 10,000+ developers. During my conversations with their engineering team, I picked up a couple of interesting insights that seem to be top of mind for them. The biggest one: Being able to customize their GraphQL gateway. Especially banks/fintechs have *incredibly* high expectations when it comes to authorization and permission management. Most API management solutions on the market can’t live up to these standards. But there’s an even bigger issue: Typically, companies adopt GraphQL because they have a lot of microservices and a large engineering team. And usually, all of these microservices follow different authorization models (RBAC, ABAC, Zanzibar… just to name a few). Originally, there was no way to customize the GraphQL gateway without additional infrastructure running in a sidecar or the ability to build customize it in a language that compiles to WebAssembly. So… we went to work and put something together: Grafbase customers can now customize their GraphQL gateway with WebAssembly. Why WebAssembly? 1/ You can customize your gateway in ANY language that compiles to WebAssembly (C++, Rust, Go, Java, Kotlin, …) 2/ Your extensions run together with the Grafbase gateway. Secure, sandboxed, and with no need for sidecar patterns or containers running in the background. 3/ There’s no need to manage another piece of infrastructure. One less thing that could break = peace of mind for you & your engineers. And with this new feature, you can customize much more than “just” your permission strategies. You can run your own custom caching strategy, modify rate limiting, etc… This feature has been on my mind for a while and I’m super happy to roll it out to the public. Excited to see what our customers are going to build with it.

19

2 Comments

If you're focused on streamlining your LLM workflows and boosting automation, this blog is for you. I compare frameworks like LangGraph, LlamaIndex, Haystack, AutoGen, and building your own agent with OpenAI, highlighting where each shines. Whether you need LangGraph’s flexibility for multi-agent systems, LlamaIndex’s workflow optimization, or Haystack’s search power, I’ve got practical insights to guide you. Key Takeaways: Start by understanding your specific needs—complexity, scalability, and integration matter. LangGraph is great for multi-agent systems, while LlamaIndex simplifies workflows. Haystack excels in search and retrieval tasks, and AutoGen is perfect for automation. Sometimes, building your own agent might be the most flexible route, though it’s more resource-intensive. https://lnkd.in/eSXsXTDX

53

2 Comments

I spend a lot of time thinking about how to design a self-serve user experience where any Capital markets team can come in and be able to automate their capital markets reporting in a couple days using our platform. If we can solve that problem, it unlocks hyper-scale because it become so easy to try and see how much better it is. As an analogy, Stripe nailed this idea in e-commerce payments. Before Stripe, to be able to integrate payments into your application was not trivial. It was a lot of conversations that engineers needed to have across a lot of implementation calls. There was so much pre-build you needed to get started. What stripe designed was this incredible self-service onboarding flow for developers. They didn't have to have an account management team come in and teach them how to start implementing stripe into their application. They could just go in, read the documentation, and start coding. The tag line in fact was “access e-commerce payments in 2 lines of code.” We are bringing that to capital markets reporting so that every lender can have enterprise grade data reconciliation and reporting up and running in a few days.

7

I may be on vacation, but that doesn't mean I can't still get excited about a new LLM model drop! Llama 3.1 just hit, and in addition to enhancements to the 8B and 70B versions, we also have a new 405B parameter version! That guy is HUGE. https://lnkd.in/g4DcKvDx #llm #llama #ai #machinelearning #genai

3

1 Comment

"It is anticipated that there will be 75% fewer employee-facing application software companies and 50% fewer categories." from Brett Queener's newsletter. The proliferation of software solutions has led to an overwhelming number of tools for employees, reducing actual productivity and mastery of their craft. Adding on top terrible UI/UX in B2B SaaS leaving users to navigate interfaces that suck. AI can simplify user interactions, understand user needs, and perform tasks autonomously, shifting the cognitive load from the user to the software. The application software market is poised for a paradigm shift driven by AI, leading to fewer but more powerful software solutions that significantly enhance user productivity. Companies that embrace this transformation will set new benchmarks for usability and efficiency, ultimately redefining what it means to be an application software company.

2

I just finished an insightful episode of Lenny's Podcast Lenny's Newsletter with Alex Komoroske! Alex’s approach to product development is all about embracing the ‘gardener mindset,’ where you nurture emergent opportunities instead of just executing a static plan. This resonated with me deeply as a product manager, as it highlights how we can allow our products to evolve naturally and create more value. Alex also discussed how AI and LLMs are fundamentally changing how we build products—making software development faster and cheaper, but not necessarily better unless we learn how to harness these tools properly! His reflections on organizational dynamics, especially his slime mold analogy, got me thinking about how we structure teams to encourage more autonomy and creativity. For anyone working in product, tech, or leadership, this is a must-listen episode. And if you're looking for book recommendations, check out "Origin of Wealth" and "The Elements of Thinking in Systems" for some systems-level inspiration. #ProductManagement #AI #EmergentThinking #Leadership #OrganizationalStructure #SystemsThinking #Innovation #LennysPodcast https://lnkd.in/exMp_aqW

4

My latest for TechCrunch: Another Black Exit! Open source compliance and security platform FOSSA has acquired developer community platform StackShare, the company confirmed to TechCrunch. Founded by Yonas Beshawred, StackShare is one of the more popular platforms for developers to discuss, track, and share the tools they use to build applications. This encompasses everything from which front-end JavaScript framework to use to which cloud provider to use for specific tasks. Read more to learn about the company and its acquisition!

61

2 Comments

I am acting this week as a Data Engineer. Why? Radix is managing a massive PayPal merchant with over 600K active subscriptions and getting 15K on average in monthly subscription trials. This base data produces secondary massive data every hour, such as failed or successful payments, refunds, and, most importantly, tracking subscriptions by status: Active, Past Due, or Cancelled. Our current data pipeline manages 95% of our customers perfectly fine, but we have some customers with massive data requirements that necessitate a more powerful data pipeline. Radix's bread and butter is data accuracy; therefore, data ingestion, processing, and storage are key; otherwise, we won't be able to assist them effectively in helping them to track and analyze their revenue metrics. Now, the issue lies in the following: as a small startup, we need to use our resources very strictly to increase our margin per customer without compromising quality. A well-implemented data pipeline can definitely help. 🔗 Data Fetching → Data Transport → Data Processing 1) AWS Fargate: Running a Python script to fetch data from the PayPal API in batches. 2) SQS: Queuing messages and storing data until it's ready for processing (perfect for batch processing). - Note: While Kafka is ideal for real-time data streaming, our PayPal APIs update every 4 hours, so it’s not necessary for our current needs. Plus, Kafka can be costly! 3) MongoDB: Storing raw data in batches. However, with gigabytes of data per fetch, it can be expensive. For maintaining speed and efficiency, AWS S3 is a better option for raw data storage. - Note: If you need to store gigabytes of raw data, AWS S3 is the best option. You can process the data from there and then store the processed data in MongoDB. This way, you will save tons of money. 4) Apache Spark: Our secret sauce! Designed to process terabytes of data in real time with the right architecture. This enables us to create vital revenue metrics like MRR, Churn, LTV, and subscription statuses. - Note: If real-time metrics are needed, processing data directly from the transport layer (SQS or Kafka) is ideal, with Kafka being the better option for streaming. Always think about reducing computing power to save costs! 5) Save the processed data, which is the bread and butter. In this case, you have several options: - Once the data has been processed by Apache Spark, you can directly save it back to MongoDB in a new collection. - Alternatively, you can save the processed data in a new database or data lake. In my case, I will use MongoDB to store both the raw and processed data. #DataEngineering #AWS #MongoDB #ApacheSpark #PayPal #DataAnalytics #Innovation #TechForGood

13

1 Comment

Really cool chart on hallucination mitigation techniques for LLMs. The most common techniques I've heard from enterprise builders are retrieval augmented generation and supervised fine tuning - both enable companies to build on top of their proprietary data and therefore attain a moat on the performance of their specific use case. Haven't seen anyone leveraging these other techniques, but if you are, would love to connect and learn more! Check out the paper here: https://lnkd.in/e8k5Accc

25

1 Comment

Check out Michael Levan and Matthew Voget chatting about how #API development is changing. They highlight a killer application of #AI (and also how you can easily start building with OpenAI's API). Jump to minutes 10 - 13 for a great snippet on the typical complexities of API dev and testing that is automated away with Blackbird from Ambassador

11

We're starting to do more cost analysis work with Depot, which is starting to highlight some fascinating things. First, our default runner is about 2x faster on average, just out of the box, and without any other integrations into Depot. The way the math works out for a basic Next.js build is quite impressive for 100 builds/day over a month: - GitHub-hosted runners would charge you about 250 hours of build time at $120/month 😬 - Depot's default runner would be 150 hours of build time at $36/month 🤯 Nearly 2x faster and 3x cheaper than GitHub, right out of the box ⚡

19

3 Comments

Shoutout to our team who worked late-nights to write this super detailed IEX Cloud <> Databento migration guide. We hope that this will bring some much-needed transparency as you compare all of the IEX alternatives out there. https://to.dbn.to/3XIgIs4

84

1 Comment

“The secret to building better business software is building a system where you can build multiple parallel business software applications that are all natively built into the same system,” Conrad said. “Now every single software vertical has like a dozen competitors, and it’s extremely crowded,” Conrad said, adding that customers prefer an approach where a software company will “build the system that sort of recombines all of this and gives people one system that combines a lot of these capabilities and wins.” Another great TechCrunch article. This horizontal approach is exactly what we're doing BuyerMatch.ai for Real Estate Brokerages. One system to do everything and one AI that understands it all. https://lnkd.in/e2JeU5ia

4

1 Comment