Iris Weinstein’s Post

A word from F5: signatures can be created quickly in a flash against new threats using artificial intelligence, Using LLM models trained to generate regular expressions, we can create signatures that can be deployed to the edge for real-time detection. #F5 #llm #API # AI https://lnkd.in/dtJMNxNa

A word from F5: signatures can be created quickly in a flash against new threats using artificial intelligence, Using LLM models trained to generate regular expressions, we can create signatures that can be deployed to the edge for real-time detection. #F5 #llm #API # AI https://lnkd.in/dtJMNxNa

What is Apple's real approach to AI? There was an amazing duality at the WWDC 2024 event: on the one hand, AI is important but was pushed only to the second part of the event. On the other hand, Apple has its own models, but there is also a connection to ChatGPT that will serve as a backup. Here is my full interpretation of Apple's developer event: https://lnkd.in/dvBqW_jP What about you? Did you like what was announced?

The business world is undergoing a revolution, with advanced AI technologies and innovative task management systems at its center. Here are some fascinating insights that can upgrade your business -> https://lnkd.in/eCvZ9YYa #innovation #moreadmin #tech #newera

At tasq.ai, we believe that excelling in AI requires addressing two major challenges: data and power consumption. #Human #guidance is essential in overcoming the data challenge, ensuring our AI solutions are both effective and efficient. Recent incidents, like Microsoft's AI mistakenly accusing a reporter, underscore the need for robust AI oversight : https://lnkd.in/dg-S-2KH Even the most successful models require a human layer to ensure responsible and ethical AI. Interested in learning how tasq.ai can help you navigate the complexities of AI with #human-#guided insights? Contact us today

AI tools are gradually gaining traction in small businesses. The key challenge lies in encouraging employees to embrace these tools and explore their potential. Read more about my thoughts on the matter on Calcalist כלכליסט https://lnkd.in/eAzSg6_7

The field of artificial intelligence (AI) is transforming our lives and has sparked a significant impact worldwide, heralding a true revolution across various sectors and industries. While the application potential of AI presents vast opportunities, its utilization also poses numerous challenges, risks, and concerns, necessitating regulation in this innovative and groundbreaking field. The European Union Parliament recently approved the world's first law for regulating artificial intelligence, known as the EU AI ACT. This law aims to govern the use of AI, prescribing requirements and imposing fines for non-compliance. So, what should you know about this new law? Dr. Eyal Brook, the partner and head of the AI practice at our firm, sheds light on these questions in the following article: https://lnkd.in/ek5VChjE

𝐓𝐡𝐞 𝐀𝐈 𝐰𝐨𝐫𝐦 !! called 𝐌𝐨𝐫𝐫𝐢𝐬 𝐈𝐈. Yes the first claimed RAG based Generative AI worm is out that works on 💥𝑨𝒅𝒗𝒆𝒓𝒔𝒊𝒂𝒍 𝒔𝒆𝒍𝒇 𝒓𝒆𝒑𝒍𝒊𝒄𝒂𝒕𝒊𝒏𝒈 𝑷𝒓𝒐𝒎𝒑𝒕💥 ⛔️The researchers have shown how the AI worm can attack a #generative #AI email assistant to steal data from emails and send spam messages—breaking some security protections in #ChatGPT, #Gemini, and #LLAVA in the process. ⏮️ 𝑨 𝒃𝒊𝒕 𝒐𝒇 𝑯𝒊𝒔𝒕𝒐𝒓𝒚 𝒂𝒃𝒐𝒖𝒕 𝑴𝒐𝒓𝒓𝒊𝒔 𝑰 ⏮️ In 1998, there were fewer than 100,000 connected machines, and most organizations that were online seemed to trust one another. There were few passwords and other hoops to jump through to get connected. Safety issues are rampant in a system like this, and Morris intended to point them out. He only meant for his worm to expose how quickly an attack could unfold, but he made a devastating coding mistake. 🪱𝗧𝗵𝗲 𝗠𝗼𝗿𝗿𝗶𝘀 𝘄𝗼𝗿𝗺 𝘄𝗮𝘀 𝗺𝗮𝗱𝗲 𝘁𝗼: ➡️ℚ𝕦𝕖𝕣𝕪 : The worm asked each computer it encountered if it already had a copy of the code. ➡️ℝ𝕖𝕤𝕡𝕠𝕟𝕕 𝕥𝕠 "ℕ𝕠." If the computer wasn't infected, the code would execute. ➡️ℝ𝕖𝕤𝕡𝕠𝕟𝕕 𝕥𝕠 "𝕐𝕖𝕤." If the computer was infected, the worm wouldn't copy. This code seems innocent, but Morris didn't want clever programmers to work around him and prompt all of their computers to respond "𝕐𝕖𝕤" to every copy. 𝑺𝒐 𝒂𝒇𝒕𝒆𝒓 𝒔𝒆𝒗𝒆𝒏 "𝒀𝒆𝒔" 𝒓𝒆𝒔𝒑𝒐𝒏𝒔𝒆𝒔, 𝒕𝒉𝒆 𝒄𝒐𝒅𝒆 𝒅𝒖𝒑𝒍𝒊𝒄𝒂𝒕𝒆𝒅 𝒂𝒏𝒚𝒘𝒂𝒚. ❌ This flaw devastated computers. Many had several versions of the code running at the same time, and performance slowed to an absolute crawl. Some systems crashed altogether. Reference: https://lnkd.in/dKJvqSHf https://lnkd.in/dEKkepFV 🚨 𝐍𝐨𝐰 𝐚𝐛𝐨𝐮𝐭 𝐌𝐨𝐫𝐫𝐢𝐬 𝐈𝐈 🚨 The worm named as Morris II is itself seems alarming. All the generative AI models works on #prompts and the same can be weaponized against it. #ComPromptMized This is a prompt that triggers the generative AI model to output, in its response, another prompt. The AI system is told to produce a set of further instructions in its replies. Let me explain it in a very simple way. 🌠𝐂𝐡𝐞𝐜𝐤 𝐭𝐡𝐞 𝐚𝐭𝐭𝐚𝐜𝐡𝐞𝐝 𝐢𝐦𝐚𝐠𝐞. Pardon my scribbling. 🚀What we need to do now is to be ready for these potential risks and start building our systems more resilient, as we intend to embed gen AI almost everywhere. ✅𝑺𝒐𝒍𝒖𝒕𝒊𝒐𝒏𝒔 𝒕𝒐 𝒆𝒙𝒑𝒍𝒐𝒓𝒆: 1️⃣Synthetic Data Generation for any queries. 2️⃣Training models using synthetic data 3️⃣Decentralization and Incentivization on data 4️⃣Governance and intervention. More on these solutions later 👋. All views are perosnal. #followformore #genai #innovation #digital #ai #meity #bfsi #openai #google #krutrim #bharatgpt #grok

I had the honor of being interviewed by Globes about the recent groundbreaking advancements from OpenAI and Google. It's fascinating to witness the rapid pace of innovation and how these giants are pushing the boundaries of what's possible with AI. Key takeaways from the interview: OpenAI's GPT 4o Launch: OpenAI introduced GPT 4o, a model that’s twice as fast and half the cost of its predecessor. With 100 million active weekly users, the goal is to make AI models even more accessible and integrated into daily life. Exciting features like video calls with ChatGPT are on the horizon! Google's Gemini Upgrade: At Google’s annual developer conference, they showcased enhancements to their AI model, Gemini, which can now handle more data and produce even more sophisticated outputs. Google also emphasized the integration of AI in their search engine and the development of advanced personal assistants. The Future of AI: Both companies are striving towards creating intelligent personal assistants that can handle a myriad of tasks seamlessly. This includes recognizing emotions in voice and eventually through video. In the interview, I mentioned: "Until now, integrating individual text segments into models was a manual process. The future model aims to provide continuous and comprehensive responses, seeing everything you see. However, it's unclear when this will become a reality." Challenging Business Models: It's intriguing to see how these companies might be disrupting their own business models. OpenAI’s move to offer GPT-4 for free aims to build a vast user base, which could convert to premium users in the future. For Google, integrating AI into search might reduce ad revenue but is essential for maintaining relevance. NVIDIA’s Crucial Role: Both OpenAI and Google acknowledge NVIDIA’s pivotal contribution with their advanced GPUs, which are essential for these AI advancements. This collaboration underscores NVIDIA's importance in the AI ecosystem. The AI landscape is incredibly dynamic, and it's a privilege to be part of this revolution. Looking forward to seeing how these technologies evolve and shape our future! Link: https://lnkd.in/dn_hFv5K