If only someone told me this before my 1st startup: 1. Validate idea first. I wasted at least 5 years building stuff nobody needed. 2. Kill your EGO. It's not about me, but the user. I must want what the user wants, not what I want. 3. Don't chaise investors, chase users, and

It's hard to predict the future of AI and the course of its development. Until recently, everyone was curious about when the "AlexNet" moment of language models would arrive. Nowadays people are talking about when the "GPT" moment of image models would arrive 😊

🚨 [New Paper] If you're involved in AI safety or jailbreaking, you don't want to miss this: Techniques from human communication now effectively breach aligned LLMs (Llama-2 Chat, GPT-3.5, GPT-4) with over 92% attack success rate. 👇🧵(1/7 - page link: chats-lab.github.io/persuasive_jai…)

The GOAT of tennis Novak Djokovic said: "35 is the new 25.” I say: “60 is the new 35.” AI research has kept me strong and healthy. AI could work wonders for you, too!

Microsoft released their annual Future of Work report and this time around it’s not about remote work, it’s about AI. (slide numbers in brackets) - Knowledge workers with ChatGPT are 37% faster, 40% higher quality but ~20% less accurate. Simple UX solutions to solve this are

For the first time ever, more than 300 innovators from #KIT came back to home for a special event. It's amazing to see how a university can be a productive incubation place for great ideas to find their way in business. We're glad to be part of the #KIT family! #kih2024

1/ Today in Science, we train a neural net from scratch through the eyes and ears of one child. The model learns to map words to visual referents, showing how grounded language learning from just one child's perspective is possible with today's AI tools. science.org/doi/10.1126/sc…

Glad to hear that #EU has chosen to follow the path for responsibility and safety. #AIAct is with us!

Today is historic as the first ever law that attempts to manage the risks and responsible use of #AI is set to be in place. The so called #AIAct of #EU will be voted in parlament in April as the last step. Then it's time to put the theory in practice and is harder than rhetoric.

Is there.....?

Here the math: Option 1: 1 Vision Pro + 5 transpacific economy flights ≈ $8k Option 2: 1 business class + 4 economy transpacific flights ≈ $8k I would absolutely take option 1.

Leak, Cheat, Repeat: Data Contamination and Evaluation Malpractices in Closed-Source LLMs paper: arxiv.org/abs/2402.03927 "we extensively document the amount of data leaked to these models during the first year after the model's release. We report that these models have been

Whoever tells you “we understand deep learning” just show them this. Fractals of the loss landscape as a function of hyperparameters even for small two layers nets. Incredible

There's a reason why people call generative AI "generative". Hallucinations are necessary, but trusting them is not.

Imagine that, Schmidhuber in 2016 talking about world models before Yann LeCun “invented” them 🤷‍♂️ More broadly, world models are old idea in cognitive science and robotics, going back decades. They are desperately important for AI and still lacking in current models, and

🔍 Concerned about #LLMs and bias? We dive into the potential bias pitfalls of Large Language Models and introduce a useful benchmark evaluation of different LLM providers. Check it out! 🚀 validaitor.com/how-unfair-are…

This is one of the most remarkable archeological discoveries of the recent past, which sheds doubts on the mainstream narrative. Must see if you visit Turkey.

First day at Crowdstrike, pushed a little update and taking the afternoon off ✌️

Cybersecurity needs more of security for itself as the risk of excessive trust to it is real: techcrunch.com/2024/07/19/fau…

Adopting ISO/IEC 42001 early can significantly benefit start-ups by: 🌱 Establishing Ethical Practices 🤝 Mitigating Risks 📜 Ensuring Regulatory Compliance #GenAI startups can hugely benefit from it. See it as a chance not as a burden!