Karl Sponholz

Navigating the Waters of Generative AI: Insights from the European Startup Landscape in 2024 🚀 Exciting times are ahead for European startups in the generative AI space! A recent study by AppliedAI Institute for Europe has thrown light on the vibrant ecosystem and the unique challenges we face. 📊 Key Findings: Germany, France, the Netherlands, and Sweden are leading the charge with the highest number of AI startups. Despite having robust innovation, European AI startups are still playing catch-up in funding compared to their US counterparts, with a stark difference in capital inflow. 🔍 Challenges Ahead: Funding: European startups have raised €2.37 billion, significantly less than the US giants. Regulatory Maze: 24% of startups feel stifled by complex regulations. Computational Constraints: Access to high-performance computing remains limited. 🛠 Strategic Moves for Startups: Visibility Boost: Step up your presence in European AI forums. Collaborative Efforts: Pool resources and knowledge with fellow innovators. Regulatory Engagement: Shape policies that foster AI innovation. 🌍 The road ahead is filled with opportunities and hurdles, but with strategic planning and collaborative efforts, we can transform challenges into stepping stones for success. Let's leverage our collective strength to navigate these waters together! 💡 Your Thoughts? I’m keen to hear how others in the AI community are tackling these challenges or if there are new opportunities on the horizon. Share your views! #AIStartups #GenerativeAI #EuropeanInnovation #TechInvestment #AIRegulation #HighPerformanceComputing #CollaborationInTech more ....>>

17

13 Kommentare

Do you want to build a new #blockchain AI app and don't know where to start? Good news! Starknet #Berlin Meetup is around the corner and you can learn everything you need to know from experts like StarkWare and Gizatech AG .     👉 23 April, 6 PM! Secure your spot! 👇 👇 👇 https://lu.ma/x5w6qfud We will kick things off with StarkWare breaking the ice, exploring the world of Starknet, and sharing the latest updates on what the Exploration team at StarkWare has been up to lately. Gizatech AG will delve into building and scaling ZKML solutions powered by the Starknet Stack. They will particularly discuss their latest development, Giza Agents — a modular and extensible framework for integrating decentralized ML into on-chain strategy and action. After that, we will wrap up the meetup with a Q&A session and networking, followed by food and refreshments. 🍕 🍻 Don't forget to secure your spot! 👇 👇 👇 https://lu.ma/x5w6qfud #Blockchain #ML #AI #Ethereum #Germany

2

Yes, there are still exciting AI news: Google's NotebookLM now turns dry factual text into lively podcast conversations. An interesting use of AI is transforming existing content—from long to short, German to English, text to speech. And now, from factual text to podcast-like banter: two people talking to each other. Google has two beta products with this feature: Illuminate, which converts scientific articles into conversations you can listen to like a podcast. Meanwhile, NotebookLM's Audio Overview does the same, because why not double the fun? (Alessandro Alviani noticed it first.) I threw my latest LinkedIn post on image creation into the AI blender. Out came a breezy five-minute chat. But beware: NotebookLM doesn't just parrot your words; it freelances with its own tidbits. And yes, there are still glitches. Did it just switch to Chinese in the end? Right now, it grabs everything in a notebook. Google's promising more control soon. From the Discord: "We're working on a way to sub-select sources and steer the audio more in the next version." Until then, the workaround is to create a new notebook and select only the desired sources or just one source. I like to recommend NotebookLM as an introduction to the world of AI tools and RAG because it quickly delivers aha-moments. Even more so now! Recently, I wrote about my kitchen cabinet: https://lnkd.in/e-bUKHBU On a more dystopian note, are we on the brink of a future where every podcast is a rambling AI-generated chat fest? Amazon had to restrict the uploading of self-published books because users were clogging the platform with too many AI-generated junk works. But if AI podcasts can save us from another voyeuristic true crime story, maybe they're not so bad.

45

#OpenAI just released #GPT-4o which enables interaction with text, audio and video in realtime, basically enabling anyone to have a Human-AI Hybrid party anywhere any time ;) Looking forward to also try out the MacOS app coming soon for a more integrated and native experience. #GenerativeAI #ConversationalAI #AI https://lnkd.in/gnFCPSB7

20

AI destroys jobs. And creates new ones. What jobs will GPT 4o create?  GPT 4o was presented yesterday. Others have reported on the functions. But what impact will the innovations have on established business models - especially on us knowledge workers from service providers and consultancies? According to a study by Goldman Sachs, up to 300 million jobs worldwide are jeopardised by generative AI. I think that's an understatement. BP, for example, says they already need "70% fewer coders from third parties to code as the AI handles most of the coding." The active assistants introduced yesterday are accelerating the revolution. The first to be affected are tutors. Last night, GPT 4o checked my son's homework and not only suggested improvements, but also praised the conditional forms he had used correctly. At the end there was a "well done" from GPT for him. The point is: what works with students works everywhere. Expert knowledge and the ability to impart this knowledge are rapidly and massively losing their value. But the change continues: the requirements for "hand-delivered knowledge" are already pricing in the capabilities of AI models - what GPT knows is the starting line of a service, no longer its destination. What follows is a "terror of excellence", as Victor Bellu quite rightly calls it: no one is allowed to go below GPT's knowledge. Those who cannot precisely articulate their added value will not survive the change. But what exactly is the added value of our knowledge compared to the machine? I think there are a few, but they need to be understood and utilised in a targeted manner. - "Data generator" This may sound a little too prosaic, but the ability to understand the value of data and integrate it into the AI engine in such a way that it creates tangible value is not recognised by every service provider.  - "Innovation Visioneer". When knowledge is ubiquitous, it's about relating things that were previously unrelated. Steve Jobs called it "Connecting the Dots". This "out-of-the-box" thinking will experience a renaissance with AI.  - "Process Engineer". Knowledge is nothing without control, or in other words: what results from knowledge, how do we take action - efficiently, sustainably, intelligently. Those who understand the "next best activity" have mastered AI.  - "Interaction consultant" - or put more simply: being human. As crazy as it sounds - the more capable the machines, the more important the human element becomes, especially in a professional environment. Empathy, attention, motivation, confirmation, closeness - a machine will soon be able to do a lot of these things too. But it will never be able to nourish us emotionally in the same way. So when will we see the first real "AI enabled Consultants"?

24

2 Kommentare