Vlad Călin’s Post

🔥 Exciting News: Llama 3.1 just released! The open source AI model you can fine-tune, distill and deploy anywhere. The latest instruction-tuned model is available in 8B, 70B and 405B versions. Designed for developers, researchers, and businesses, Llama 3.1 is equipped to drive the next wave of generative AI innovation. Choose from real-time inference or batch inference services. - Inference Choose from real-time inference or batch inference services. Download model weights to further optimize cost per token. - Fine-tune, Distill & Deploy Adapt for your application, improve with synthetic data and deploy on-prem or in the cloud. - RAG & Tool Use Use Llama system components and extend the model using zero shot tool use and RAG to build agentic behaviors. - Synthetic Data Generation Leverage 405B high quality data to improve specialized models for specific use cases.

AI is taking a massive leap forward this year. Here are 5 important trends that we are currently observing: 1. More Realistic Expectations: Businesses now have a refined understanding of AI-powered solutions. It's all about integration, not revolution. 2. Multimodal AI: Expect advances in generative AI that can handle multiple types of data inputs, providing more intuitive and versatile applications. 3. Small(er) Language Models and Open Source Advancements: With smaller models becoming more powerful, AI is becoming more accessible and explainable. 4. GPU Shortages and Cloud Costs: As cloud computing costs increase, the trend towards more compact, efficient models grows. 5. Customized Local Models and Data Pipelines: As the playing field levels, competitive advantage will be driven by proprietary data pipelines. 

Generative AI market is set to experience a significant surge in 2024 📈 , with high adoption rates. But how can we navigate this wave of innovation effectively 🤔 ? Will Grannis, VP and CTO at Google Cloud, suggests an approach based on three pillars: 𝐬𝐮𝐬𝐭𝐚𝐢𝐧𝐚𝐛𝐥𝐞 𝐜𝐨𝐬𝐭𝐬 💰 , 𝐚𝐜𝐜𝐞𝐬𝐬 🚪 , and 𝐭𝐫𝐮𝐬𝐭 𝐚𝐧𝐝 𝐬𝐞𝐜𝐮𝐫𝐢𝐭𝐲 🔐. In this post, I would like to focus on the first pillar - economics and energy - as I believe it highlights the crucial role of 𝐅𝐢𝐧𝐎𝐩𝐬 👨💻 👩💻 in the era of Generative AI adoption. To ensure you're maximising your investment in Generative AI, it's essential to ask the right questions. Here are some thought-starters: 💬 Are you equipped with the foundational capabilities to maximize your Generative AI investment? 💬 Have you established robust financial controls to manage Generative AI costs? 💬 Can you identify and maximize the business value of your Generative AI use cases? 💬 Have you implemented a cost model and governance process for onboarding new Generative AI use cases? FinOps can bring immense benefits to Generative AI projects. It provides a framework for managing the costs of cloud computing ☁ 📲 , including those associated with AI workloads. By implementing FinOps practices, organisations can manage these costs and maximise the value of their AI investments 💲 . It's not just about cost control, but about enabling innovation, accelerating time to market, and driving business value. Remember, designing with cost-aware architecture is key 👷♂️ 👷♀️ 🔑 #GenerativeAI #FinOps #AIAdoption #AIInvestment #CostManagement #ValueRealization #CloudComputing #AIWorkloads #Innovation #BusinessValue #AIJourney

More AI Momentum from Google Cloud Next 2024! Google's Gemini 1.5Pro Public preview and AI Hypercomputer are set to revolutionize the industry. With AI-powered Google Workspace and various use cases, customers are innovating with Google AI like never before. The competition is heating up, but the good news is that there is always a GenAI available for every use case. Let's connect to discuss how GenAI can enhance your business and explore its various use cases. #AI #GoogleCloudNext2024 #GenAI

Fantastic recognition for the Seaco AI sprint team from Google & Google Cloud, Ministry of Digital Development and Information, IMDA and Singapore Economic Development Board (EDB). Seaco just picked up an 'Innovation award' for the 2024 Google Cloud AI Trailblazer program! We are incredibly proud to support Singapore's NAIS2.0 initiative through action and are living up to our AI Verify Foundation pledge! Of course we are not stopping here, the future is really exciting as we transform our industry vertical. Since embarking on the Trailblazer program we now have another three MVPs built, implemented across our hyperscaler platforms as working full production use cases. As we build momentum and add more features to these live use cases we will continue to experiment, discover and collaborate with our business peers to drive value and continue to solve the most difficult industry challenges. Those that have heard me speak about AI will know, I am a huge fan of 'building your own' AI solutions as opposed to leaving up to the SaaS vendors who may or may not develop use cases that you need / want. Though building vs buying is definitely harder route to take as a team, it has proven to drive great value to us as a company as it allows us to; 1. Own our data throughout its journey 2. Control the path and location that the Data takes in transit and at rest 3. Create a secure AI platform approach to integrate all managed platforms 4. Own the roadmap, patterns, modelling ourselves e2e 5. Experiment, make our own mistakes, learn and grow as a team I wanted to thank our partners Soo Ming Poh, Cherie Kweh, Lidya Salim, Bharath Yaragatti. My IT team that delivered; Lawrence Coelho, Dani Satheesan, Karthikeyan Rajarathinam. Most of all I would like to give a big shout out to our business sponsors; Vincent Lo, Zenn Sim, Tat Shnenn Lim, Brian Rysz, Felix Fu, Eng Eng Chen, Mark Goh, Rohit Saxena without your support and your teams collaboration this would not have been possible. As we continue to invest in our people and platforms I am excited about the future of AI for us and the positive impact it will have across the entire maritime industry.

2024 was predicted to be the year where we see Gen AI move from experimentation to production and after 3 packed days at Google NEXT, that definitely seems to be coming to fruition. With Gen AI - Theory and experimentation is one thing, but production and impact is quite another. To demonstrate this, and the difference Vertex [our integrated end to end AI platform] makes, we’ve put together a summary of 101 use cases across industries, building for production with Google Cloud. It’s thought, only 10% of businesses have Gen AI in production today and by 2026 this number is forecasted to be 80%. Where are you on your journey? #Next24 #GenAI

The big question, long-term, is whether demand for AI services will be worth the investment for each company — the value proposition may seem clear but the customer ROI models are yet to be defined relative to CAPEX & OPEX ‼️

Very interesting report and must read. I liked 3 realities authors predicted. There are three possible realities we expect to see in the foreseeable future on who will capture the most value in this model layer fight: - In Reality 1, AI models will be commoditized, similar to compute or oil, becoming essential assets for global business operations. The ultimate value in the AI ecosystem will be captured by compute and cloud service providers, marketplaces, and applications — not by the models themselves. - Reality 2 suggests that a few AI Model Giants, backed by Big Tech strategics or corporate VCs, will dominate the foundational model ecosystem, following a trend similar to the Cloud Wars. - Reality 3 envisions an AI model economy resembling the diverse and popular potato chip market, where various model companies thrive due to unique use cases like modalities, performance, latency, cost, and security, allowing for their survival. Trends are changing every moment as we browse. Let us watch out! #AI #ArtificialIntelligence #FuturePredictions #TechTrends

📌 According to Global Market Insights, the AI as a Service market size was valued at $6.4 billion in 2022 and is anticipated to grow at a CAGR of 25% between 2023 and 2032. ✨ Join us in this latest episode of Cloud Cover and hear thought leaders from IBM, Lenovo and Amazon Web Services (AWS) explore how AI as a Service is joining the growing list of “as a service” tech solutions that business customers are deploying to: — increase customer value — remain cutting edge and — differentiate their products and services in the marketplace. When: 2/21 Time: 12 EST   🔑 Register Here:  https://lnkd.in/eFkDQsfG   🔔 Subscribe to the Channel Here: https://lnkd.in/et3YD9NS

Great read from Sequoia Capital . They really nailed the business model shift from #cloud to #ai. - Cloud companies targeted the software profit pool. AI companies target the services profit pool. - Cloud companies sold software ($ / seat). AI companies sell work ($ / outcome) - Cloud companies liked to go bottoms-up, with frictionless distribution. AI companies are increasingly going top-down, with high-touch, high-trust delivery models.