Oleg Ciubotaru

Back to Rust for Heimdahl.xyz's Solana integration 🛠️ After some side projects, finally putting Rust into real production. Two key reasons: 1. Solana's official SDK - no compromises with third-party libs 2. Rich ecosystem of Rust code to adapt and build upon The compiler's strictness makes perfect sense when you're building infrastructure that has to just work. Successfully streaming Solana events through the same WebSocket as EVM chains. Clean architecture paying off - new chains shouldn't mean new complexity. Next up: Cross-chain USDT flows tracking. One API to track them all. Try it out: https://lnkd.in/dF8T7jqQ Website: https://heimdahl.xyz p.s. When the Rust compiler approves your code, you know it means business. #RustLang #Solana #Infrastructure

1

We probably have more cross-project collaboration at The Gnar Company than most software agencies. Any pull requests that go up get posted on a pull request channel across the whole team. So even though most of our engineers are fully dedicated to a single project, they’re still seeing code from other work and getting their own code reviewed by others with a fresh perspective. Plus, every few weeks we have “Show the Codes” event. If there’s interest in a certain subject or a particularly interesting project, we throw some code up on the screen during lunch and talk about it. It’s normal to learn and grow based on reps at the individual engineer or even project level, but we think we can get better even faster by creating opportunities to learn at the company level. #softwaredevelopment #codereview #showthecode

12

4 Comments

AI code banned in open source projects? ⛔ Yes, but not because of its reliability or safety, but because of its "tainted" origin (that is, undefined copyright situation). And in case you are wondering: that also applies to any copy&paste from stackoverflow. #oss #generativeai #aiprogramming https://lnkd.in/etwzTSvF

7

It's finally here. https://lnkd.in/eZviiVPX Today I am happy to share the first public version of our Kalavai client (totally free, forever), a tool that helps you build an AI cluster from your everyday devices. Our first use case is distributed LLM deployment, and we hope to expand this with the help of the community (check our templates hub https://lnkd.in/e_DVBaAt) Help us reach as many hearts as possible. How? - Show us your love by starring our git repo (https://lnkd.in/eZviiVPX) - Say hi in our community (https://lnkd.in/eirwNc9i) - Try our tool and give us feedback! - Reply here and tell us your challenges when working in AI

52

5 Comments

The conventional advice for building horizontal products is to choose a single vertical and expand one by one 👀 Yet within the first 6 months of launching our SDK generator, we built support for 7 languages: Python, TypeScript, Terraform, C#, Ruby, Java, and PHP. With only 4 FT engineers at the time, was this the right decision for the business? Here’s what we learned: *️⃣ Building so many language targets in a short period of time really pressure tested our generation capabilities. We built a platform that was flexible and easy to develop on because we HAD to. *️⃣ We figured out how to access expertise from language communities that we were non-native to. For example, we’re not native PHP developers, so we’re working with people like 🤓 Steve McDougall who know the languages inside & out to keep us on the right path. *️⃣ Perhaps most importantly, we listened to our customers. Companies in different industries need different languages. C# is very important in Fintech. Most people are not interested in a Terraform provider. But for anyone whose API is considered infrastructure by their users, Terraform is critical. Focusing on breadth over depth drove early momentum for us and accelerated us into customer conversations. Each language took a good amount of time from a small team. But rapid experimentation helped us quickly validate which industries had specific language demands as well as which industries needed a bit more education on the value of SDKs. Was every language we built worth it? Head to our platform to start generating SDKs and decide for yourself! Link in the comments 👇 #sdks #langauges #api #pmf

16

3 Comments

So far, I am not impressed with the early drafts of the post-open source license. It can put your business for 1% of revenue - which most major businesses try to avoid. And it doesn't even touch yet how they will split that 1% between contributors - which is not a problem we know how to solve fairly. Oh, and you cannot use it in your software products, SaaS, or shrink-wrapped - as that would put your customers into the same "give us 1% of your revenue" category. https://lnkd.in/gXvjAeYX

7

2 Comments

Founders, it’s ok to roll your eyes when a VC asks “what’s your technical moat?” My take: technical moats no longer exist in software. It used to be that you could launch a product with a technically capable team and enjoy a two to three-year advantage. But now… there are better engineers than ever before, more platforms that can even the playing field, and AI that is helping everyone build quality software, fast. A differentiated product and 10x team matters, but your competitor might have these as well. The moats that make a difference now: • Your shipping velocity • Your network effects • Your GTM motion • Your community • Your brand • Your story • Your data • Your ROI • Your customer love • Your positioning. How are you different? What other moats have you seen in this era of more technical parity? #founders #venturecapital #categorycreation Felicis

280

29 Comments

I had the chance to play with Claude 3.5 Sonnet this weekend and integrate it into my novel drafting scripts. Some notes: * The output is good. Better than Gemini Pro for my use case. I had to do only very minor edits. That saved a ton of time. See also attached sample chapter, courtesy of Claude. * Thanks to the overall quality and the big context window, I had to send fewer queries overall to the LLM to generate the text. I estimate that I can create a good-quality draft for less than 10 cents / 1000 words. Meaning, I could create the first draft of a novel (assuming 100k words) for under 10$. Very exciting!

8

2 Comments

Currently, rollups are fragmented, leading to poor UX and disjointed liquidity. Here the first demo of Javelin by our portco NodeKit With Javelin, users can synchronously send messages and move funds across #rollups on NodeKit within a single 2s block.

11

The dumbest part of the whole post: "The best part? He spent $0 in paid marketing." Most of this is just fluff in order to drive signups to this guy's "strategy nerds" newsletter...and it works. While this guerilla marketing framework is indeed reproducible for many businesses and entrepreneurs, don't expect the same level of results and certainly don't expect it in the same amount of time (4 months). Most startups and small businesses have tighter capital restrictions and smaller teams to execute on this to achieve the same level of results. Maybe it's true Gas spent $0 on paid ads, but the whole guerilla marketing operation didn't cost $0. Keep in mind, this founder is flush with capital after a $100M acquisition from Facebook. With that amount of money, he can then hire a team of developers to build Gas and a team of guerilla and social media marketers to execute on the strategy. All of that costs A LOT OF MONEY. So maybe it's true he didn't spend money on paid ads, but he probably spent hundreds of thousands if not millions on a team to execute this guerilla marketing strategy. This didn't cost $0.

7

Marc Andreessen from Andreessen Horowitz said "Software is eating the world". Wouldn't that make everything in the world basically a massive amount of "IF THEN" statements? And software really is just a bunch of if/then code. If the body gets a cut, then start clotting and send white blood cells to prevent infection. If it's Sunday, then take out trash for Monday's pickup etc. So, if we can decipher how the brain determines the "then" part of "if" situations...then we can program machines to think like humans. #ai #activeinference

2

10 Comments

🤖 Droidcon NY Recap 🗽 My talk video is now available! The presentation dives deep into the current pain points and unique challenges of mobile observability, with a special focus on solving how to capture screen representations in Jetpack Compose apps. Also, come read about our first conference experience! https://lnkd.in/gBBDJihh

17

1 Comment

Substrate is live! After nearly a year of development, really honing the developer experience and performance tuning a very unique runtime environment, we're finally releasing what we feel is by far the best way to run multi-inference workloads (e.g. agents, multimodal text, json, image, audio, embedding pipelines, etc). Substrate lets you use a ergonomic SDK (we really obsessed over the abstractions and typeaheads) to describe precise and complex relationships between AI tasks. And—importantly—it's connected to a backend specifically designed to run these graph workloads as efficiently as possible (optimized models, auto batching, insanely fast inter-node RPCs). Given the current state of "AI" models, we really believe that this pattern is the most effective for building capable, reliable, legible, extensible features with AI that go far beyond chatbots. We need both heuristic models like LLMs as well as normal symbolic computation, and a flexible/scalable way to coordinate these two modalities. We've already seen organizations like Substack building some super interesting pipelines that they can easily understand and iterate on, that just work at scale. It's still early in a much more ambitious vision for us, but we think this V1 is not only ready but already a meaningful improvement to many implementations. Reflecting on the last year, I feel so much gratitude for the team that I'm working with: Ben Guo, Liam Griffiths, Rishabh Parikh, 💭 Kyle Pitzen, Chris S., (& Hanwen Wu soon!). With this launch, we're announcing an $8M Series Seed led by Lightspeed (Nnamdi Iregbulem+Guru Chahal) with participation from many others. Feeling a ton of appreciation for the folks that took a bet on us or have lent their time and advice: South Park Commons (Ruchi Sanghvi+Aditya Agarwal), Craft Ventures, Red Swan Ventures (Andy Dunn), Lorimer Ventures, Chris Best, Jairaj Sethi, iqram and the immigrant groove, Andrew Kortina, Guillermo Rauch, Immad Akhund, Sam Lessin, William Gaybrick, Gokul Rajaram, Jana Messerschmidt, Shreyas Doshi, Matthew Hartman, Mike Cohen, Sabrina Hahn, Anne Lee Skates, Shreyans Bhansali, Julian Connor, Namrata Patel + many many others. https://lnkd.in/eHq8KUkA

128

9 Comments

How do open source projects track interesting links? I want to keep a repository of links we come across while building SlateDB. Sometimes we include them in our RFCs, but not everything has an RFC. And some posts are just neat but not related to current work.

2

3 Comments

Things we shipped last week at Riza: - Support for sending input files along with code for execution - A way to add HTTP authentication secrets to requests from your code, without exposing the secrets to your code - New releases of all supported client libraries (Python, node.js, Go) Links and more info available from the changelog: https://riza.io/changelog

22

The market for third-party CDN services, which I define as the delivery of video and small/large objects, was about $5 billion in 2023. (Full blog post: https://lnkd.in/ekFZqX9y) This figure does not include revenue for services tied to security (WAF, DDoS, etc.), compute, hardware, and storage and excludes delivery in China. The number excludes any company with its own DIY CDN, like Netflix and YouTube, that doesn’t sell its infrastructure as a service to other companies. The figure also excludes revenue from vendors offering online video platforms, including Brightcove, Uplynk, AWS Media Services, Quickplay, Deltatre, Mediakind, Sports Radar, Cognizant, Endeavor Streaming, StreamAMG, Bedrock Streaming, etc., which combined saw revenue of about $1.8 billion in 2023. Based on what I know, AWS Media Services is the largest of the group, with about $450 million in revenue in 2023. While reports produced by research firms say the CDN market was multiple times larger than my number, they are wrong. Full stop. Do not buy their reports. Akamai, Fastly, Edgio, CDN77, and others disclose enough details tied to their delivery revenue that we have numbers to form a good foundation for market sizing. The market is not as big as many suggest, and those writing about the size of the CDN space don’t define it and include non-CDN services like transcoding, hardware, hosting, storage and transit in their numbers. Full blog post: https://lnkd.in/ekFZqX9y Listen to my special CDN-focused podcast from July 2024 for more details and updates regarding 4K, low/ultra-low latency, CMCD, DIY (TikTok, Disney), Open Caching, pricing and multi-CDN. https://lnkd.in/epzXSMVA Vendors mentioned: Akamai Technologies, Amazon Web Services (AWS), Edgio, Fastly, Cloudflare, CDN77, Medianova CDN, MainStreaming, Qwilt, Wowza, Ateme, Broadpeak, Gcore, Netskrt Systems Inc., Varnish Software, Vecima Networks Inc., Velocix.

666

39 Comments

Thanks for sharing. This is very insightful, I was waiting for when something like this would come up as there is a lot of buzz about AI, and the people who depend on it the most are not experienced software engineers. This is scary because these users will not know when they are replicating/introducing bugs into there environments, or even worse exposing themselves to vulnerabilities. While AI, may help us work faster, it should not been seen as a replacement to system analysis, requirements gathering, system design, code reviews and VAPT Testing. These interactions are very much human dependent and remain important to a successful product development/roadmap. They can not be replaced by AI just yet. Finally, for engineers, if you have to use AI, make sure it is on technology you have extensive experience and fundamental knowledge on. That way you will know when you are fed the wrong code.

1

1 Comment

Following the excellent example Rows.com, here's another from another hot company 🚀 folk (from renowned French studio Hexa, neé eFounders, with a $9M seed from Accel, Scott Belsky, Arash Ferdowsi, Nir Eyal, Nicolas Dessaigne and more) and their great UI... What I love about this one = smart empty states A lot of products are now using "cards" as a pattern when a user first lands into the product, to encourage them to take key actions But folk does a couple really smart things that I haven't always seen: 1️⃣ Sets priority Horizontally stacked cards often don't have this, and offer equally weighted CTAs for all the cards... but users want to make fewer decisions ⬇️ 👉 Folk makes the first CTA filled and contrasting with the others, so it's even easier to get started. 2️⃣ Generates data for me Often products require new users to do a bunch of work or commit to importing data before there is any reward or value. This creates friction and means more people miss the aha moment 🙈 👉 Folk offers "sample contacts" to use instead of my data, so I can get to see the value of a CRM without committing This reduces waiting time and prevents me bouncing Thanks Simo for the submission -- great example to other #founders to submit your user onboarding or in-app flows if you want some recognition 🔦 Even if you think its not the best, share an in-app flow to Chameleon's User Onboarding Olympics to earn cash and be in the pot for our 3 x $300 gold medals 🥇 We're all learning together 😇 Click here to upload a screenshot now: chameleon.io/go/olympics 🔵🟡⚫️🟢🔴 Or please repost ♻️ or comment 💬 to help your network find this great opportunity! I regularly share examples of good UI/UX and best practices for user onboarding and PLG, so follow me 🔔 (Pulkit) for more of that in your feed 👋

24

4 Comments