AI is not after your job, probably!

Along with: Tesla’s FSD v12 Elevates – How Smart Are We Talking?

Hey there, 

I hope you liked our first edition of AI Emergence - a newsletter covering AI while it is Emerging :)

If you are new here, we are building Human Intelligence about Artificial Intelligence while AI is Emerging!

What would be the format? Every week, we will break the newsletter into the following sections:

  • The Input - All about recent developments in AI

  • The Tools - Interesting finds and launches

  • The Algorithm - Resources for learning

  • The Output - Our reflection 

  • Question to ponder before we meet next!

Table of Contents

Is AI coming for your Job? Probably not, but it’s not that simple

The IMF SDNs reveal that up to 60% of jobs in developed countries are exposed to AI disruptions. While AI might boost productivity for half of these roles, it could also replace human-performed tasks in the other half, possibly affecting wages and employment rates.

But is it the complete picture? Quantifying AI's broader societal impact isn't straightforward. Adding to this, MIT's study shows only 23% of vision-related tasks might be economically viable for AI automation due to high setup and running costs.

So, is AI the job-snatcher it's often portrayed as? Not quite. The focus should be on using AI as a productivity tool while ensuring the workforce adapts alongside.

Meta is advancing into a new era with the development of Llama 3, their upcoming generative AI model. This strategic move aims to weave artificial general intelligence into their products, bringing them closer to human-like functionality. 

Meanwhile, Meta is realigning its AI Research unit, now led by Chris Cox, to better integrate research with product innovation. Complementing this shift, Meta's significant investment in GPUs (it plans to procure 350,000 Nvidia H100) reflects its commitment to advancing in the realm of sophisticated AI technologies.

This is probably the first time an organization has shared the size of infrastructure required to train these LLMs and obviously, the numbers are huge. At current prices of H100 - this would mean a few Billion dollars only for GPUs. Add the connectors, cost of people, electricity, and all other expenses and we get to a sense of the cost of building this technology. Reminds me of what my father paid to get a x386 in India in the early 90s! (source)

The 2024 World Economic Forum (WEF) concluded this week and there was one thing common across all days of the event: AI.

Prominent figures, including Sam Altman, Li Qiang, Satya Nadella, and many others, participated in the 54th Annual Meeting. Here are the most important takeaways:

  • Sam Altman's Take on AI's Future: Altman emphasized the escalating energy needs of AI technologies, advocating for sustainable solutions like nuclear fusion to meet these demands. (watch video)

  • Satya Nadella's Economic Outlook: Nadella, with Microsoft's foray into generative AI, shared a positive outlook on AI's potential to rejuvenate the global economy, highlighting its transformative capabilities. (watch video)

  • Global Leaders on AI Ethics: Discussions with leaders like Li Qiang and Ursula von der Leyen focused on AI's ethical implications, underscoring the need for responsible AI development and balanced governance to navigate its societal impacts effectively. (watch video)

Tesla's latest FSD Beta v12 is a game-changer, with a unified neural network. This innovative shift replaces a massive 300,000 lines of traditional C++ code, making the vehicle's vision and behavior fully reliant on AI. (source)

The Next Wave of AI Gadgets: Rabbit & Perplexity Collaboration

Rabbit R1's partnership with Perplexity AI marks a pivotal moment in AI gadgets. This collaboration brings to life an AI-driven gadget that connects seamlessly with services, enabling it to tap into Perplexity's advanced 'answer engine' for real-time, accurate information. (source)

Stability AI Launches Compact 1.6B Language Model

Stability AI, the pioneer in image-generative AI, has introduced Stable LM 1.6B, its most compact model to date. This model, supporting seven languages, is designed to broaden the generative AI landscape for developers globally.

Despite its smaller size, Stable LM 1.6B impressively outperforms other models in its class including Microsoft’s Phi-2 (2.7B), TinyLlama 1.1B, and Falcon 1B. and even some larger ones including Stability AI’s own earlier Stable LM 3B model.

However, like any small language model, it faces challenges such as higher hallucination rates and potential language issues. The launch marks a significant step in making AI more accessible and versatile. (source)

Lights, Camera, Text: ByteDance's MagicVideo-V2 Spells a New Era in Video Creation

ByteDance, the parent company of TikTok, has unveiled a groundbreaking text-to-video generation model named MagicVideo-V2. This model introduces significant advancements in the realm of visual content creation, outperforming leading industry models with its high aesthetic quality and fidelity. It features a multi-stage architecture that integrates a text-to-image model, a video motion generator, a reference image embedding module, and a frame interpolation module. (source)

Altman envisions a future where AI technologies require a specialized semiconductor supply chain. To achieve this vision, he aims to create a network of AI chip factories, departing from the conventional approach of depending on existing foundries. This involves engaging in discussions with potential investors, with a primary goal of securing a significant amount of funding. (source)

  • CodiumAI has launched AlphaCodium, a groundbreaking AI code-generation tool that outperforms DeepMind's AlphaCode by employing innovative flow engineering to enhance coding accuracy significantly. (source)

  • Microsoft and MasterCard are introducing AI tools like Copilot and Small Business AI, designed to help small businesses enhance their operations and growth through advanced AI technology. (source)

  • Google Chrome's latest update introduces three AI-powered features: custom generative AI themes, automatic tab grouping, and an enhanced AI-driven writing assistant, all aimed at enriching the user's browsing experience. (source)

  • Here is an interesting no-code / low-code platform for building AI assistants.

  • If you're interested in the future of AI, don't miss this video. It's a captivating conversation between tech leaders at the World Economic Forum, exploring the exponential advancements in Generative AI and the evolving landscape of risks and opportunities in this groundbreaking field.

  • In the latest episode of 'Leading with Data,' I had an engaging conversation with Thomas Wolf, Co-Founder of Hugging Face, about the development of the famous Transformers library, how he got into AI and what is the top of the mind opportunities for him.

  • The article from Analytics Vidhya provides a detailed roadmap for mastering Generative AI in 2024, outlining a progressive learning path from basic user interaction to advanced research and development in the field.

  • Article from Ethan Mollick on the impact of AI - A very good summary of the impact AI has created over the last 12 months (source)

  • The article from Analytics Vidhya discusses the top seven tools available for interacting with PDF documents. These tools transform the way users engage with PDFs, allowing for dynamic conversations, real-time summarizations, and enhanced comprehension.

I was talking to a few people who attended Davos and CES in the last few weeks. All of them felt a fatigue with AI being mentioned everywhere. Every conversation and every product seemed to be dominated by AI.

At Davos, the questions were more about building trust and security. There was also a clear acknowledgment that the current Energy demands for these AI models look unsustainable.

At CES on the other side, there were some crazy products - but the flying car (source) was something that had everyone’s attention!

Last week I asked you how the rollout of Microsoft Copilot would impact your work and Organisation. I got some very interesting responses. By and large, most of you felt that it would increase productivity, but also has a danger of creating a lot of junk documents, and mail flying around!

I couldn’t agree more. Also - I still prefer Beautiful.ai over PowerPoint presentations! Interestingly, there is a Copilot Pro subscription, which I have not tried yet.

Question of the week

For this week, I want to hear any new tools (with or without AI) that you have come across and would love to share them with me.

How do you rate this issue of AI Emergence?

Would love to hear your thoughts

Login or Subscribe to participate in polls.

Reply

or to participate.