Alex Yakunin

🚨 Looking to Buy or Invest in Real Estate? We’re offering 10 Test Users 6 𝐦𝐨𝐧𝐭𝐡𝐬 𝐨𝐟 𝐟𝐫𝐞𝐞 𝐚𝐜𝐜𝐞𝐬𝐬 𝐭𝐨 𝐏𝐫𝐨𝐩𝐞𝐫𝐭𝐢𝐮𝐦! 🚨 My partner Vasily Safronov and I, Nikita, are experienced data scientists with CTO backgrounds, and we noticed a big gap in the real estate investment market: many people have the desire and funds to invest but lack clear insights on where and when to do it. So we created Propertium — a platform built to provide high-quality data and insights to make property investment simple, transparent, and smart. 🔍 Why Propertium Is the Right Choice for Your Real Estate Investments  • 𝐒𝐞𝐚𝐫𝐜𝐡 𝐏𝐫𝐨𝐩𝐞𝐫𝐭𝐢𝐞𝐬 𝐛𝐲 𝐑𝐎𝐈: Our advanced AI model calculates real ROI, offering insights into your potential earnings. Want to know how much you could make renting out an apartment in Barcelona? Propertium provides clear estimates to ensure your investment aligns with your financial goals.  • 𝐄𝐱𝐩𝐚𝐧𝐬𝐢𝐯𝐞 𝐄𝐮𝐫𝐨𝐩𝐞𝐚𝐧 𝐂𝐨𝐯𝐞𝐫𝐚𝐠𝐞: Access property data and investment insights across Europe, from Bulgaria to Portugal, ensuring you have the widest range of opportunities to explore.  • 𝐇𝐨𝐭 𝐎𝐟𝐟𝐞𝐫𝐬: Be the first to know about new listings that match your investment criteria, so you’re always ahead of the game. Whether you’re a seasoned real estate investor or just starting out, Propertium is built to empower you with precise, easy-to-use tools that offer an "x-ray vision" into property details, helping you make informed decisions with confidence. 🚀 Become one of the first 10 users of our platform. Send us a message by November 20th and we will explain all details 🔥

11

6 Comments

Thrilled to enable 💪 Bedrock - llama 3.2 cross region inferencing - https://lnkd.in/dad-phUS s/o Steffen Ridderbusch (+4 more updates 👇) 💸 Azure AI rerank endpoint cost tracking s/o Emerson Gomes 🛠️ Logging - response redaction on `turn_off_message_logging` s/o Ilari Scheinin 🔨 Cohere - support cost override for cohere rerank models s/o Jerome Roussin ⚡️ Proxy - fix returned headers 

8

Belated Post, however, Project Prometheus participated in the 10th Goedware Game Jam; while we didn't win any significant category, the surplus of positive reviews on the Jam's Page has done wonders for the morale to work on other projects. The experience gained from being forced to work under a sacrosanct time limit has been tremendous. It has given me not only technological experience but also more familiarization with myself as an individual. Published Jan 15, 2024; ~four days of development. Linked, the game is open for playing -> https://lnkd.in/eh3cKmaJ

1

On-device capabilities are the future of many AI use cases. Privacy, cost, offline functionality, and reduced latency will all play crucial roles. Check out our latest blog on how to run Small Language Models (SLMs) and build RAG pipelines on devices. #SLMs #RAG #AI #GenAI

28

Ilya Sutskever (co-founder of OpenAI and SSI) thinks AI scaling has hit a wall and we're back in the age of discovery—finding the next big thing. Fair point; but is this something new, or alarming? From the first artificial neuron in 1943 to today's LLM, AI has come a long way. And we're still in the early days—with a long road ahead toward achieving AGI. There have been breakthroughs that redefined possibilities while introducing even bigger challenges. Every time it seems like AI is hitting a wall, there comes another breakthrough. Current models excel at pattern recognition, but true reasoning? We're not there yet. Does it mean we've hit a wall with the transformer architecture? Not necessarily. Scaling laws still do work. However, since they follow a logarithmic pattern, it means you need exponentially more resources (compute, data, memory) to achieve marginal gains. It's getting harder and harder to reach the next milestone, given the available resources. Hence the roadblocks: → Architectural limits: Transformer-based architectures have inherent constraints. → Efficiency hurdles: Bigger isn’t always better. Multimodal models, with varied sizes and capabilities, could be the future. → Physical constraints: Power, compute, infrastructure, and hardware require significant innovation. → Data integration: Real-time data incorporation remains a critical challenge. What will AI look like in a few years? The pace of development has been exponential—anything seems possible. Excited to see where this journey leads. #AI #DeepLearning #ML #AGI

55

1 Comment

We benchmarked Apollo vs WunderGraph vs Mesh vs Grafbase GraphQL Federation gateways. Here are the results (and my take on the benchmarking): CONTEXT: Last year, we launched Grafbase GraphQL Federation (built in Rust) from scratch. To see how it performs compared to others, we’ve started benchmarking it against other providers. Here’s the full benchmarking setup: - All gateways had the same CPU and RAM resources, limited by Docker. - We repeatedly executed a complex query using the load-testing tool K6, with a constant and ramping VU scenario. - The final response content is 88KiB big, merged from several subgraph responses. Now to the interesting part… RESULTS: Throughput: - Grafbase Gateway: 1502 req/sec - Cosmo Router: 867 req/sec - Apollo Router: 434 req/sec - Mesh: 300 req/sec Latency: - Grafbase Gateway: 57.9 ms - Apollo Router: 92.4 ms - Mesh: 127.2 ms - Cosmo Router: 308.7 ms (We’ve benchmarked more than just throughput & latency, the full blog post is linked in the comments below) MY TAKE: First, I’m happy to see that all the hard work we put into the Grafbase Gateway paid off. I believe it’s mainly due to two things: 1/ We betted on Rust to build the Grafbase Gateway right from the start – and that’s still paying dividends today. Some other core design choices also contributed to this result. 2/ As a company, we’re striving for a culture of performance. This goes beyond our engineering team, but definitely reflects in product performance. But what does this result mean from a business perspective? 1/ First, great performance helps you reduce your infrastructure footprint. For enterprise companies, we’re talking about billions of requests per minute (yes, minute). These seemingly small differences quickly turn into hundreds of thousands of dollars in infrastructure costs. 2/ It’s a better foundation for scaling. This is especially relevant for large enterprises. If you expect heavy load on your systems (because you want to scale), your foundation needs to be rocksolid. At this level, there’s no room for small performance issues. 3/ Better user experience Above all, better performance means a better experience for your users. I think I don’t need to elaborate on this. The effects of great (or not so great) user experience can be felt in *every* aspect of your company. – To be honest, I was also impressed by some of the results. The sustained latency vs. throughput metric showed that you can keep throwing requests at the Grafbase Gateway and get guaranteed great latency. If you’re interested in the full benchmark, check out the link in the comments. All that’s left to say: Kudos to our amazing engineers at Grafbase. – PS - We tried to be objective. If you feel like we should benchmark other metrics or see ways to improve - comment below, file a complaint or open up a PR. PPS - If you encounter a case where Grafbase doesn’t provide this level of performance, we’d be happy to look into it for you.

40

8 Comments

I just built an Agreement Redliner Agent in <5 mins with Lyzr AI Agent API Studio. Here's what happened 👇 One of our partners got in touch, mentioning they had a client who needed a agreement redliner agent. While I could share a somewhat similar case study, the partner wanted something even closer to their requirements. That's when I thought, why not try building one? With Lyzr Agent API Studio (https://agent.lyzr.ai/) , I quickly designed an agent with the 'no-code wizard' that can, 1. Check the agreement for any objectionable clauses 2. Redline the clause 3. Share the reason for redlining I gave the agent five key clauses to look at. (described in the app) And voila. I just built an agent, and also an app for the agent, in under 5 minutes. And no code. And not even OpenAI o1. We used GPT-4o. Lyzr's agents have in-built ReAct framework that helps in superior reasoning and execution. And Lyzr agents automatically enhances agent prompts by applying 'chain of thought' structure. How I built the agent - https://lnkd.in/eEpYXMVi Try the agent - https://lnkd.in/eKJY5Ykn Build your own agent - https://agent.lyzr.ai/ If there's a process you're thinking of automating, know that Lyzr lets you do so without chasing your IT team. Got a complex or critical business process or job role to automate, speak to us. https://lnkd.in/eh6ih-9q

87

4 Comments

Ramblr’s Data Engine now supports automated #SceneGraph generation! This powerful feature bridges the gap between visual scenes and human language, enabling advanced vision-language tasks and seamlessly injecting video context into #LLM prompts. Drago Guggiana Nilo walks you through the functionality and shows you how to get started. Get your free account at https://ramblr.ai and try automated #SceneGraph generation on Ramblr.ai®'s platform Stay tuned for more - We are constantly improving our  #DataEngine enabling advanced #VideoUnderstanding. #AR #MR #VideoIntelligence #ContexualAI #AIforThePhysicalWorld 

10

FOR CODERS ONLY - Rust #genai video tutorial Rust #genai - One Client, One API, Multi-AI Providers Currently natively supports: #openai #ollama #gemini #anthropic #cohere https://lnkd.in/gi37Cnic #rustlang #generativeAI

8

2 Comments

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

Ilya Sutskever's new startup, Safe Super Intelligence, just raised $1B (rumored at $5B valuation - source https://lnkd.in/gSvCAbr9). Based on the company name and the OpenAI story, this seems to be about making #LLMs safer. I can understand Ilya and his team commanding high investment, but I suspect that there is more than just adding a layer of safety on top of current LLM scaling to command such a high investment. Will be interesting to see what ideas they come up with in their models. #AGI #ASI #SSI

1

Launching on Product Hunt? Don't miss these tips from Simona O’Neill She has on ProductHunt: → 1k+ points 🏅 → ~250-day streak 🌞 → 600 followers 👥 Discover how she's managed the launch of Canvid: - Gained 1000+ website visits in just days - Converted 20% of visitors - Built lasting community relationships Plus, learn the unexpected benefits that came AFTER their launch (applicable only to AI products). Curious? Check out our full interview: https://lnkd.in/dVY9xy7j

20

7 Comments

Just implemented the Hermitage test for VeloxDB, and all tests passed! 🎉 Hermitage tests evaluate how databases manage transaction isolation by checking for known anomalies. Passing all tests shows that VeloxDB adheres to serializable isolation (the highest level of isolation, which prevents all types of transaction anomalies). You can check out the implementation here: https://lnkd.in/dA_QHw6s

73

4 Comments

🗞️ Ilya Sutskever the former Chief Scientist of OpenAI has just launched a new company, Safe Superintelligence Inc. (SSI) 🤖 🧠✨ Co-founded with Daniel Gross, who oversaw Apple’s AI and search efforts, and Daniel Levy, formerly of OpenAI SSI aims to tackle AI safety head-on, ensuring advancements in AI capabilities are matched by robust safety measures. 🛡️ 🔍 Key Insights: 👩‍🔬 Mission-Driven: SSI is entirely focused on AI safety and capabilities, solving these challenges through revolutionary engineering and scientific breakthroughs. 🔐 Unique Approach: By prioritizing safety ahead of capabilities, SSI plans to scale AI advancements peacefully and securely. 💰 For-Profit Model: Unlike OpenAI's initial non-profit stance, SSI is designed as a for-profit entity, aiming to attract significant capital. 🗣️ Sutskever commented "Our singular focus means no distraction by management overhead or product cycles, and our business model means safety, security, and progress are all insulated from short-term commercial pressures." 🌍 Offices are located in Palo Alto and Tel Aviv, and they are actively recruiting top technical talent. For more details, check out the full article: TechCrunch 👇 #AI #TechNews #Innovation #SafetyFirst #SSI

17

5 Comments

Step-by-step walk thru of new Groq native modules for Make with all new native JSON mode, plus OpenAI whisper for translations + transcriptions. It makes automating AI so much faster and a little bit easier. https://lnkd.in/g6bHyKVe

2

This week we’re launching 10BY10 - a ten week series hosted by Avii, Agentiiv’s digital person. Every week Avii will be summarizing a complex report or whitepaper on AI and The Future of Work, and delivering a concise 2 minute video summary of the main takeaways and a more detailed written review with a link to the original report. The full workflow, including the summarization, video script writing and of course Avii herself is AI augmented. The objective is to share meaningful value through distilled insights that you can digest in under 10 minutes, and give you the breadcrumbs to learn more if you’d like to access the full report. This week we’ll start off with a comprehensive look at how marketers are adapting with the "2024 State of Marketing AI Report from Marketing AI Institute”. Check out the detailed written summary in Avii’s blog here:  https://lnkd.in/ghS2jA2v

80

26 Comments

Once users get used to a certain system performance, it becomes the standard. This is problematic with systems like Prometheus, which has been steadily optimized to a local maximum over the last 10 years. Problematic because we know there are architectural changes that can get it into the area of a much better local optimum, but it would require first starting over and living with a less performant system for some time. Is there a name for this kind of performance dilemma?

24

8 Comments