Shai Alon

🚨 Elevate your SOC to the next level with the power of AI, alert modeling, and proactive response! 🔍 Discover how these cutting-edge technologies can revolutionize your threat detection and incident response capabilities in our latest blog post. 🔒💻 #CybersecurityInnovation #SOCOptimization

4

"We are entering the new SaaS era, moving from Software-as-a-Service to Service-as-a-Software." - this recent statement by Andreessen Horowitz is as profound as Marc Andreessen’s 2011 assertion that “software is eating the world.” It’s time for software companies to shift their focus from delivering tools and features to delivering outcomes. The current state of AI allows companies to do that, addressing the complexity, underutilization, and lack of customization that plague SaaS. There’s an interesting parallel here to the shift from on-premise to the cloud. Back in the day, organizations had to maintain both infrastructure and software to achieve their desired outcomes. Cloud solutions allowed companies to outsource infrastructure management. Now, with Service-as-a-Software, the utilization of software can be outsourced too, enabling organizations to focus entirely on outcomes. This is a new and compelling model for tech entrepreneurs. Instead of building tools and features, they can focus on delivering outcomes using AI and other technologies to lower costs. This approach allows startups to generate revenue quickly and achieve profitability from the outset, without the immediate need for venture capital. For organizations, this model eliminates the need for extensive software installation, configuration, training, and onboarding. They can simply prescribe the outcomes they need, have service providers deliver those outcomes directly , and start reaping the rewards right away, all while keeping costs low. Service-as-a-Software is not just a future trend—it’s the present opportunity that businesses must embrace. This is the operating thesis at my new venture, and we are doubling, even tripling down on it. I’m happy to share our learnings with anyone interested. Do you believe the same? I’d love to hear your thoughts.

21

What is common between the Olympic Games and Product Management? Watching the incredible achievements of the Israeli team at the Olympic Games inspired me and got me thinking about the similarities between the journeys of athletes and product teams. Here are a few:  **Vision and goal setting: clear, ambitious goals and strategic planning drive both athletes and product teams forward.  **Continuous improvement: just as athletes undergo rigorous training, product teams continually refine their products based on feedback.  **Teamwork: success in both fields relies on collaboration—athletes with their coaches and teammates, and product managers with cross-functional teams. **Adaptability: both must adjust to uncertainty, whether it's on the field or in the market. **Resilience: overcoming challenges and striving for long-term success. Have more to share? Feel free to add in the comments. #productmanagement #olympics

17

I’ll admit it—I really enjoy building PoCs and demos! It’s one of my favorite parts of the day-to-day as a co-founder. Recently, I put together a demo on reading Avro messages on MSK and saving them as Parquet files in S3. Since I wouldn’t make you sit through the whole 5-minute video, here’s a quick 1-minute “trailer” to give you the gist! 🎬 Here’s my workflow for making these demo videos. If you’re in DevRel, I’d love to hear any tips on making this process even smoother! Step 1: Set up a working PoC/demo. No photoshop please. Step 2: Think about the flow and key highlights (no need to write them down). Step 3: Record the screens with QuickTime—audio optional. Step 4: Watch the rough cut, jot down what you’ll say (incomplete sentences and typos are 💯 fine). Step 5: Let AI refine the text, but check it over carefully. AI likes to get (too) creative sometimes. Step 6: Use TTS (OpenAI or Eleven Labs) to generate the audio. Step 7: Edit the screensharing video and match up the audio—a process that somehow takes hours, if not days! Step 8: There are always changes. You just need to record some screens and re-generate the audio for the new requirements. Hopefully you don't need to edit the entire video again. Would love any feedback or tricks you use to streamline your own workflow! 😊

27

9 תגובות

Wix launch of AI creation capabilities for its mobile app builder. Through a conversational AI chat experience, Wix users describe the intent of the app, resulting in a professional, uniquely designed, and fully customizable, application. This new solution changes how users can create and modify their business apps with AI technology for iOS and Android and without code. The mobile app builder guides users through a conversational AI chat experience to understand their intent and goals in creating a new native app. Using the information from the chat, the AI generates a fully branded mobile app for the desired business. Users can customize their app from the Mobile App Editor to make additional changes according to their needs, such as sections, widgets, designs, themes, and more. Users can easily see a preview of the app on their mobile device before going live on the App Store and Google Play.

99

1 תגובה

Haya Axelrod Stern will present at the Haya Data conference (Haya^2!) how Natural Intelligence leverages GenAI to democratize the creation of complex analytical flows. Don't miss this opportunity to learn from an expert! #GenAI #Innovation #hayadata

23

I am more and more excited about Local ML on Apple Silicon. In particular, Apple Silicon + Core ML is starting to be a very nice stack and they've improved the Developer Experience significantly at the latest WWDC. Pedro Cuenca and Christopher Fleetwood wrote a blogpost on HF about the main DX and performance improvements announced at WWDC 2024: - Swift Tensor, a much simpler Tensor API (numpy, but for Swift) - Stateful Buffers (pretty much required for modern ML) - new built-in quantization techniques The post then describes how to convert Mistral 7B to Core ML, and the end result is: You have a 7B parameter model running at 30+ tokens/second using less than 4GB of memory on your Mac. 🔥🔥🔥🔥

1,534

81 תגובות

Calling all Hebrew-speaking Flink users! In this on-demand recording, Daniel Cohen of Yotpo shares how his team leveraged Flink and achieved: ⭐ Cost Savings ⭐ Improved SLA ⏱️ ⭐ Performance Optimization Watch the video and get ready to master Flink on Kubernetes! 👉 https://ow.ly/U51O50RqbVo #flink #kubernetes #apacheflink #devops #bigdata

12

Think of the biggest challenge you’re currently facing as you grow your business. Now, imagine having a supportive community of seasoned founders giving you guidance and advice to overcome this very challenge! This is no pipe dream – Google Cloud has built the Founder’s Story community just for you ↓ Listen to Merav Bahat, CEO & Co-Founder of Dazz, share her biggest challenges, how she overcame them, and how Google Cloud has been integral to her business’s success. #CloudForIsrael #FoundersStory

2

Looking forward to this insides! Agenda: Roughly 20 min presentation + time for Q&A. In case you did not sign up yet, here is the link to the sign up page: https://lnkd.in/djqdgf8f #ServiceDesign #NonGovernmentalOrgs #OnlineSummit #LearnFromExperience

1

AI startup idea: train agents. Some say being a real estate agent is like driving a taxi—temporary and anyone can do it. Agents' turnover is huge, corporate coaches are exhausted, and real customers don't appreciate being guinea pigs 😅 Got another idea for a side project based on an AI Real Estate Assistant: Young Agent Simulator An AI assistant like Placy is good at talking with clients. She understands how they act. She can play the role of a buyer or seller to help train new agents, making the whole process cheap and easy. Plus, she can send over recommendations for students and the agency’s management about each trainee. This saves money and reduces customer stress Good idea: 💡 insightful bulb   Nobody will use it: 😁 funny face Related post. AI startup idea: instant survey biz  https://lnkd.in/dQUFvxuF @AI_assistant #startup_idea #Placy 

26

1 תגובה

TL;DR; use LLMs. Start to use them. Start to understand how to use them. Start to think how they can help you in your life. Start to think how you can use it, integrate it in your business. If you don't then you might find after a while that others did and outpaced you as a person and your company. This is not a joke. Longer version: It's been a while since I posted anything on LinkedIn but a development in the last few months is basically encouraging me to inform you about something you already read about, used and discussed for hours with your friends, family and colleagues: The rise of AI / Large Language Models and how useful they can be (or not). To keep it short ; first versions of LLMs were bad. They were compressed versions of the internet which could give some answers, often looking 'good' but being bad. A year ago this improved and it lingered around an IQ of say 85 for some of the models from 2023. Not good enough to threaten anyone but the lowest administrative and programming jobs, but ... it had some uses for some people and some companies. People started to use it, to get better results, more personalized results than your regular Google Search and after learning how to actually ask a good question to a machine, the results started to become better and better. Fast forward to about 2 months ago. OpenAI released version "o1" and this model is pretty large (there are models which use up to 2 TB of VRAM - yeah that's a lot, even their quantized versions can easily be 400 GB of VRAM - but I digress). But size doesn't matter if you don't use it properly and oh boy, does "o1" use its size properly. Some people have compared the intelligence of this model to be around 120-130. It knows of many subjects from math, physics, chemistry to reasoning and 'speak' a dozen languages without a problem, with a response time measured in seconds. Try to find a single person who can do all that for you. 24/7. The downside is that it is still AI/LLMs and there is hallucination, it will forget after a while what you said earlier in the conversation (the context window is relatively small - about 128k words) but the thing is ... this is a big jump from say 85 to 120 IQ in just a year or so. And 3 years ago the AI wasn't able to coherently answer questions at all. So what is so interesting ? In a year from now we might see more and models approach this very high level of precise and accurate knowledge which tops 99% of humans. Doesn't mean it can do what you are doing - history and previous experiences are still extremely important for doing a job right, but "it" will be able to automate a lot of simple processes which have a lot of parameters but each parameter is rather simple. Just like planes are already flying and soon also taking off mostly on their own. And those planes don't even use AI. It's just code. Is this a threat? Sure. If you let it be. Instead see the LLMs as a very versatile tool, a swiss army knife. If you don't, others will.

5

3 תגובות

As part of our talk with Anh Chu, Solution Architect at Microsoft, we discussed the maturity levels of copilots that we see SaaS companies building to infuse their products and offerings with Generative AI. 💡 Key Takeaways: 🔹 Starting with Fundamental Functionalities 🔹 Progressing to Integrated Assistance 🔹 Reaching More Autonomy

22

4 תגובות

Description: Welcome to The Unstoppable Product Pioneer podcast, where Tamar Stolz, a seasoned expert in product strategy and SaaS innovation, guides you through the intricate world of product management. With a passion for empowering product managers at mid-size SaaS companies, Tamar shares practical insights on achieving strategic alignment, driving growth, and fostering cross-functional collaboration. In each episode, Tamar dives deep into the strategies and frameworks that successful companies use to streamline processes, align goals, and boost market impact. Whether you're looking to optimize workflows, strengthen your team's leadership, or sharpen your product management skills, Tamar delivers actionable advice to help you stay ahead in the competitive SaaS landscape. What You’ll Learn: How to align your product strategy with your business goals Best practices for product management coaching and mentorship Proven methods to streamline team processes and enhance collaboration Leadership development tips to build strong, high-performing teams Join Tamar as she explores the strategies that help organizations accelerate development, improve product-market fit, and increase revenue. Subscribe now for expert discussions, tutorials, and more to empower your team and drive product success. Subscribe to The Unstoppable Product Pioneer and start building products that matter! Connect with Tamar: 🌐 Website: uxpressed.com 📧 Email: Tamarstolz@uxpressed.com 📅 Book a Call: calendly.com/tamarstolz 🔔 Newsletter: tstolz.substack.com 📺 YouTube Channel: @TheUnstoppableProductPioneer

9

5 תגובות

𝗛𝗼𝘄 𝗢𝗹𝗹𝗮𝗺𝗮 𝗶𝘀 𝗖𝗵𝗮𝗻𝗴𝗶𝗻𝗴 𝗢𝗽𝗲𝗻 𝗦𝗼𝘂𝗿𝗰𝗲 𝗔𝗜, 𝗟𝗟𝗠, 𝗚𝗲𝗻𝗔𝗜 𝗪𝗼𝗿𝗹𝗱! 🌟 🔍 𝗪𝗵𝗮𝘁 𝗶𝘀 𝗢𝗹𝗹𝗮𝗺𝗮? Ollama is a tool that helps people use open source LLMs more easily. It offers a simple interface and easy-to-use features that let anyone work with these advanced AI models. 💡 𝗛𝗼𝘄 𝗢𝗹𝗹𝗮𝗺𝗮 𝗠𝗮𝗸𝗲𝘀 𝗧𝗵𝗶𝗻𝗴𝘀 𝗘𝗮𝘀𝘆: 1. Simple Interface: Easy to use, even if you’re not a tech expert. 2. Quick Integration: Add AI to your projects with minimal effort. 3. Scalability: Handle more tasks without extra hassle. 🌐 𝗥𝗲𝗮𝗹-𝗪𝗼𝗿𝗹𝗱 𝗘𝘅𝗮𝗺𝗽𝗹𝗲: --> Imagine a small business wanting to improve customer support with AI. --> Before Ollama, setting this up was require significant technical expertise and resources. --> Now, with Ollama, they can quickly create a smart chatbot to answer customer questions, improving service and freeing up staff for other tasks. 𝗖𝗵𝗲𝗰𝗸 𝗼𝘂𝘁 𝘁𝗵𝗶𝘀 𝗰𝗵𝗮𝘁𝗯𝗼𝘁 𝗶𝗺𝗮𝗴𝗲 𝗳𝗿𝗼𝗺 𝗮 𝗽𝗿𝗼𝗷𝗲𝗰𝘁 𝗜'𝗺 𝘄𝗼𝗿𝗸𝗶𝗻𝗴 𝗼𝗻. 𝗜 𝘂𝘀𝗲𝗱 𝘁𝗵𝗲 𝗠𝗶𝗰𝗿𝗼𝘀𝗼𝗳𝘁 𝗽𝗼𝘄𝗲𝗿𝗲𝗱 𝗣𝗵𝗶-𝟯 𝗟𝗟𝗠 𝗺𝗼𝗱𝗲𝗹 𝘄𝗶𝘁𝗵 𝗢𝗹𝗹𝗮𝗺𝗮 𝗮𝗻𝗱 𝗟𝗮𝗻𝗴𝗖𝗵𝗮𝗶𝗻. Follow Basavaraj N Hirebidari for more AI, LLM, ML, GenAI. #ai #ml #llm #chatbot #businessgrowth #opensource #ollama

3

June was a huge month for the Athina AI (YC W23) team. The Highlights: - Launched Athina IDE publicly and got a record number of new sign ups and subscriptions - Shipped an incredible set of new features that customers are already using HEAVILY - 🚀 Over 400 PRs merged (some serious eng. velocity for such a small team) - ⚡ Introduced Dynamic Columns to run prompts, execute code, make API calls, and more so you can work interactively with your datasets - 🥷 Introduced a code execution evaluation so you can create your own evals with python functions - 🧮 Athina IDE got a lot more spreadsheet features like hiding, resizing, adding and deleting columns, editing cells, filter, search, sort, etc. - 🗂️ Prompt management got a MAJOR upgrade with version control, SDK / API access and organization features. - 💼 We added support for custom models and improved our on-prem deployment process for enterprise customers + lots more (see the changelog in the comments) (We had to introduce a changelog because we could not send customer updates as fast as we could ship new features 😅) The best part of a new month is zooming out and seeing what you accomplished last month. The progress really stacks up. 🚀 cc Himanshu Bamoria Vivek Aditya Akshat Gupta Akhil Bisht

93

5 תגובות

🚀 𝗥𝗲𝗮𝗱𝘆 𝘁𝗼 𝗿𝗮𝗶𝘀𝗲 𝗺𝗶𝗹𝗹𝗶𝗼𝗻𝘀 𝗳𝗼𝗿 𝘆𝗼𝘂𝗿 𝘀𝘁𝗮𝗿𝘁𝘂𝗽? 𝗛𝗲𝗿𝗲'𝘀 𝗮 𝗰𝗼𝗺𝗽𝗿𝗲𝗵𝗲𝗻𝘀𝗶𝘃𝗲 𝗴𝘂𝗶𝗱𝗲 𝗳𝗿𝗼𝗺 𝗛𝘂𝘀𝘁𝗹𝗲 𝗙𝘂𝗻𝗱 𝗩𝗖! 👇 1️⃣ 𝗪𝗵𝗼 𝗶𝘀 𝗛𝘂𝘀𝘁𝗹𝗲 𝗙𝘂𝗻𝗱? Hustle Fund is a VC firm that invests in pre-seed software startups. Their mission? Democratizing wealth through startups by catalyzing capital, knowledge, and networks. 🌍💡 2️⃣ 𝗪𝗵𝘆 𝗥𝗮𝗶𝘀𝗲 𝗠𝗼𝗻𝗲𝘆? Not every business should raise money. If you can bootstrap, do it! But for high-growth tech startups aiming for big exits or IPOs, fundraising is key. 💸✨ 3️⃣ 𝗙𝘂𝗻𝗱𝗿𝗮𝗶𝘀𝗶𝗻𝗴 𝗦𝘁𝗮𝗴𝗲𝘀 𝗘𝘅𝗽𝗹𝗮𝗶𝗻𝗲𝗱 - 𝗣𝗿𝗲-𝘀𝗲𝗲𝗱: Just getting started, aiming to build an MVP. 🛠️ - 𝗦𝗲𝗲𝗱: Validating market demand with real user feedback. 📈 - 𝗣𝗼𝘀𝘁-𝘀𝗲𝗲𝗱: Achieving milestones for Series A. 🌱 - 𝗦𝗲𝗿𝗶𝗲𝘀 𝗔: Found product-market fit, ready to scale. 🚀 4️⃣ 𝗔𝗻𝗴𝗲𝗹 𝗜𝗻𝘃𝗲𝘀𝘁𝗼𝗿𝘀 𝘃𝘀. 𝗩𝗲𝗻𝘁𝘂𝗿𝗲 𝗖𝗮𝗽𝗶𝘁𝗮𝗹𝗶𝘀𝘁𝘀 - 𝗔𝗻𝗴𝗲𝗹𝘀: Individual investors, smaller checks ($1k-$25k), quick decisions. Ideal for early stages. 🕊️ - 𝗩𝗖𝘀: Larger checks ($50k+), can participate in later rounds, well-connected. Perfect for scaling. 📊 5️⃣ 𝗖𝗿𝗲𝗮𝘁𝗶𝗻𝗴 𝗮 𝗞𝗶𝗹𝗹𝗲𝗿 𝗣𝗶𝘁𝗰𝗵 𝗗𝗲𝗰𝗸 𝗘𝘀𝘀𝗲𝗻𝘁𝗶𝗮𝗹𝘀: - 𝗧𝗲𝗮𝗺: Showcase your all-star team. 🌟 - 𝗣𝗿𝗼𝗯𝗹𝗲𝗺: Identify a big, solvable problem. 🔍 - 𝗦𝗼𝗹𝘂𝘁𝗶𝗼𝗻: Present a compelling solution. 💡 - 𝗠𝗮𝗿𝗸𝗲𝘁: Prove there's a huge market. 📊 - 𝗧𝗿𝗮𝗰𝘁𝗶𝗼𝗻: Show real user or revenue growth. 🚀 6️⃣ 𝗕𝘂𝗶𝗹𝗱𝗶𝗻𝗴 𝗜𝗻𝘃𝗲𝘀𝘁𝗼𝗿 𝗥𝗲𝗹𝗮𝘁𝗶𝗼𝗻𝘀𝗵𝗶𝗽𝘀 Start early and be genuine. Build relationships with potential investors long before you need their money. Authenticity is key. 🤝 7️⃣ 𝗛𝗮𝗻𝗱𝗹𝗶𝗻𝗴 𝗜𝗻𝘃𝗲𝘀𝘁𝗼𝗿 𝗠𝗲𝗲𝘁𝗶𝗻𝗴𝘀 - 𝗙𝗶𝗿𝘀𝘁 𝗠𝗲𝗲𝘁𝗶𝗻𝗴: Nail your elevator pitch and be ready to answer tough questions. 🗣️ - 𝗙𝗼𝗹𝗹𝗼𝘄-𝗨𝗽: Immediate follow-up is crucial. Show your commitment. ✉️ 8️⃣ 𝗗𝗲𝗮𝗹𝗶𝗻𝗴 𝘄𝗶𝘁𝗵 𝗥𝗲𝗷𝗲𝗰𝘁𝗶𝗼𝗻 Not hearing back? Ask for feedback. Keep refining your pitch and approach. Persistence pays off. 💪 9️⃣ 𝗣𝗼𝘀𝘁-𝗙𝘂𝗻𝗱𝗿𝗮𝗶𝘀𝗶𝗻𝗴 𝗦𝘁𝗲𝗽𝘀 - 𝗖𝗲𝗹𝗲𝗯𝗿𝗮𝘁𝗲: Share the success with your team. 🎉 - 𝗘𝘅𝗲𝗰𝘂𝘁𝗲: Focus on delivering results and scaling your business. 📈 🔟 𝗙𝗶𝗻𝗮𝗹 𝗧𝗶𝗽𝘀 - Be transparent and honest with your investors. - Understand your valuation and dilution. - Always be aware of current market conditions. For more insights and details, check out the full guide by Hustle Fund VC! 🌟 #isapinvestments www.isap.investments www.starthub.agency iSAP Group Investments ltd StartHub.Agency

3