Jesse Anton

Anthropic’s latest generative AI model, Claude 3.5 Sonnet, is outperforming industry giants like OpenAI’s GPT-4o on key benchmarks after its release. This new model is reported as Anthropic’s “most intelligent model yet,” excelling in reasoning and coding tasks while running twice as fast and at a fifth of the cost of its predecessor, Claude 3 Opus. Claude 3.5 Sonnet was released last Thursday and has quickly climbed to the top of league tables, including the LMSYS Chatbot Arena for coding and hard prompts, ranking just behind GPT-4o overall. It is part of a three-model family, combining features of the Haiku and the Opus models, designed for efficient, high-throughput tasks. #ai

15

2 Kommentare

It’s true: the Swiss don’t miss on #opensource While several European countries have been making gains with the adoption of open source software (OSS), Switzerland’s EMBAG legislation is a milestone – taking OSS to a new level. What are the risks and rewards? On CMS Critic, I talked to Mathias Bolt Lesniak – an open source evangelist affiliated with TYPO3 GmbH and the Open Website Alliance. We’ve spoken before about the opportunities and challenges with open source in the EU, and he has some keen observations about Switzerland’s bold step to mandate OSS for its public sector agencies. According to Mathias, Switzerland’s bold leadership on open source reflects a culture - and a long history – of democratic ideals. “Democracy is what I usually point to when I talk about the positive things open source can bring into countries.” It’s a great discussion about the evolution of OSS – and where things might be headed from here. Check it out: https://lnkd.in/eCy4HTdX

16

1 Kommentar

From France to the EU: A Test-and-Expand Approach to EU AI Regulation: Reflecting on Thomas Jefferson's insight that "Laws and institutions must go hand in hand with the progress of the human mind," we see this dynamic unfold in the context of artificial intelligence (AI) regulation within countries that are not leading in AI innovation, such as those in the European Union (EU). These countries often expedite AI regulation, potentially as a mechanism to mitigate their lack of direct financial benefit from AI advancements by controlling how AI is deployed. This approach underscores the notion that regulations frequently trail behind technological innovations and need continual updates to keep pace with technological progress. Drawing from Justice Louis Brandeis's concept (New State Ice Co. v. Liebmann ) of states as "laboratories" of democracy, a more nuanced approach might serve the EU better than its current path toward an EU-wide AI Act. By allowing individual member states, like France, to first test and refine AI policies at a more localized level, the EU could better assess the impact and efficacy of such regulations. This method would enable the EU to implement policies that are proven to work on a smaller scale before scaling them up across all member states, ensuring that regulations are both effective and adaptable to the fast-paced evolution of AI technologies. Such a strategy would promote a more organic integration of regulations and technological development, potentially fostering an environment where innovation can thrive alongside necessary safeguards. My full article can be found here: https://lnkd.in/gbqZnqPw Find more post like these in my AI & Policy blog here https://lnkd.in/gXaqQfHa * These views are not legal advice but business opinion based on reading some English text written by a set of brilliant people. ** Content © Aditya Mohan. All Rights Reserved. Robometircs, Amelia, Living Interface and Skive it are trademarks of Skive it, Inc. The content is meant for human readers only under 17 U.S. Code § 106. Access, learning, analysis or reproduction by Artificial Intelligence (AI) of any form directly or indirectly, including but not limited to AI Agents, LLMs, Foundation Models, content scrapers is prohibited. 

1

1 Kommentar

Happening Today at the BTOE Summit! "How AI Can Help Drive Decision-Making in Operations" Today from 2:15 PM to 2:45 PM. Discover how AI is changing the way we work and make decisions. Don't miss it! See you there! #BTOESummit #OperationalExcellence #GlobalLocal #BusinessTransformation #ActiveOps #BTOESInsights BTOES Insights

6

1 Kommentar

🙋‍♂️ Update about AI Regulations!! 👀 (My fav. topic 😅) The European Union has announced the 'AI Act', establishing the world's first comprehensive regulatory framework for artificial intelligence!!! (Finally 💁‍♂️) It seems, this legislation, developed by April 2021, its goal to ensure that AI systems used in the EU are safe, keep fundamental human rights, and let EU countries align with EU values.. 'The Act' also introduces strict regulations for high-risk AI applications, such as biometric surveillance and emotion recognition, and requires AI-generated content, even including systems like ChatGPT 👊 to be 💥 TRANSPERENTLY LABELED!!! 💥 This positions the EU as a global leader in AI governance, setting a standard for responsible AI development and usage.. The AI Act categorizes AI systems by risk level and mandates specific regulatory measures accordingly. It includes provisions for general-purpose AI and foundational models, requiring developers to provide detailed information about the capabilities and limitations of their AI systems. The legislation emphasizes human oversight for high-risk applications and transparency regarding the training data used in generative AI models, protecting intellectual property rights and fostering public trust​.. I understand that the AI Act action, aims to harmonize AI regulations across member countries, transitioning from voluntary to mandatory compliance. But also mentioned that the Act is expected to take full effect in 2025, allowing time for adaptation and compliance etc. . This regulatory milestone is anticipated to influence global AI policies and encourage other regions to adopt similar standards​.. 🤷‍♂️ And, well, we'll see that. I am not positive on that.. Still US, China and Russia will do whatever they want to do in terms of getting the lead in that technology..

14

4 Kommentare

#ArtificialIntelligence is aimed at reducing costs and bringing productivity in the ecosystem. the #EUAIAct puts an huge onus on regulators and auditors aiming at controlling the effects of AI. While the legal fraternity would encourage such regulations, this would put burden on technology companies and regulators. in longer run, citizens would not realize the full benefits of this new age technology. https://lnkd.in/gGgq_G-u

21

1 Kommentar

The term "bureaucracy" dates back to the 18th century, combining the French bureau (desk) and the Greek kratos (power)—essentially, "political power from the writing desk." This system helped scale our built environment and infrastructure over the last 200 years by organizing records, first in drawers and now in digital files. However, while bureaucracy has its benefits, it has also given rise to "red tape"—slowing down processes, creating silos, and leaving regulations lagging behind technology. In today’s complex, interconnected world, this can stifle innovation. That’s why I’m thrilled about the announcement today: The Rt Hon Peter Kyle has launched a new Regulatory Innovation Office to fast-track public access to emerging technologies and fundamentally rethink how bureaucracy works. A fantastic initiative to help drive economic growth! Congratulations to the Department for Science, Innovation and Technology #connected #innovation #growth #datasharing Rt Hon Rachel Reeves Peter Kyle Erika Lewis Dr Alison Vincent, CBE FREng Mark Enzer Sarah Hayes https://lnkd.in/epfgJyhp

18

2 Kommentare

Digital scent trails in #LLMs Google’s #Gemini has implemented and deployed watermarking while it is probable that OpenAI and Anthropic have it available but not fully released. And as of Oct 2024, “#HuggingFace has now made the Google watermarking tech available to all 150,000 of its hosted LLMs.” A lot of people, I think, falsely believe from the nature of the intimate 1:1 communication medium, that their outputs are imbued secretive. LifeArchitect suggests otherwise for future use cases of watermarking (with anyone who holds the master access key). * Creation of searchable databases linking individuals to all #AI-generated text they produce, enabling unprecedented surveillance * Employers using watermark detection to terminate employees for unauthorized AI use, even for basic tasks * Academic institutions retroactively revoking degrees upon discovering AI-assisted work * Insurance and other companies scanning customer communications to deny claims based on AI-detected language patterns * Government agencies scanning public discourse to identify and track AI-generated political speech However, the technology is not yet robust for prime time. “But there are ways in which the signal can be removed — by paraphrasing or translating the LLM’s output, for example, or asking another LLM to rewrite it. And a watermark once removed is not really a watermark.” #watermarking #generativeai #outputtracing #outputsurveillance https://lnkd.in/gE-qS-BJ https://lnkd.in/gQRstvyj

12

** Tomorrow Thursday 10th, 3pm CET (9am US ET) on the IAMONES stage: Rob Byrne, Heiko Klarl and myself will debate on burning questions such * What will AI/LLM change in the IAM Consulting Revenue model? What will go up, what will go down, and what will disappear? * Predictions on the IAM TCO? How much will go down tx to LLMs? * Will we still need SAP wizards in our IAM projects? Registration link in the first comment. Hey Ho let's go. Make Identity fly.

50

1 Kommentar

#Web3 #SelfSovereignIdentity #ZeroKnowledgeProofs #DecentralizedIdentifier Understanding Zero-Knowledge Proofs in Digital Identity Systems : as we look towards the future, the adoption of ZKP-based digital identity solutions is poised to increase across various sectors. The integration of ZKPs with emerging technologies like blockchain and decentralized identifiers (DIDs) opens up new possibilities for secure and decentralized identity management. ZKPs have the potential to become a fundamental building block of secure and privacy-preserving digital interactions. By enabling trustless verification and selective disclosure, ZKPs can revolutionize the way we authenticate, authorize, and share personal information in the digital world. The future of ZKPs in digital identity is driven by continuous innovation and research. As new techniques and protocols are developed, the capabilities of ZKPs will expand, unlocking new use cases and applications beyond digital identity. Ongoing exploration and experimentation will push the boundaries of what is possible with ZKPs, driving the development of more advanced and efficient solutions. As the digital landscape evolves, ZKPs will play a crucial role in shaping a more secure, private, and trust-based future for digital interactions. https://lnkd.in/ehTVgrHV

2

3 Kommentare

🔑The Next Web's feature of Terra Quantum AG today shows how we make it easy for developers and security professionals to protect sensitive data against current and future threats. Our new TQ42 Cryptography has a user-friendly unified API for post-quantum (PQ) and quantum-resistant algorithms, essential security features such as key generation and management functions, and a scalable architecture. 📚 This open-source library is the latest component of our quantum-as-a-service ecosystem, TQ42. It can help enterprises make data transmission, storage, and authentication more secure. ➡ Read the full article here: https://lnkd.in/gAkAZGRT ➡ Link to the repository: https://lnkd.in/gPjwPn-U #QuantumIsNow #cryptography #quantumtech #quantumcommunication #quantum

53

4 Kommentare

This new law requires all public bodies to disclose the source code of software developed by or for them unless third-party rights or security concerns prevent it. This "public money, public code" approach aims to enhance government operations' transparency, security, and efficiency. #Mattermost #opensource #missioncritical

4

Training advanced #AI #models like #OpenAI ’s #ChatGPT and #Google ’s #Gemini Ultra requires millions of dollars, with costs escalating rapidly. As computational demands increase, the expenses for the computing power necessary to train them are soaring. In response, AI companies are rethinking how they train generative AI systems. In many cases, AI companies are finding new solutions for #training #language #models to combat rising costs. These include a number of approaches, such as creating smaller models that are designed to perform specific tasks. Other companies are experimenting with creating their own, synthetic data to feed into AI systems. However, a clear breakthrough is yet to be seen. Today, AI models using synthetic data have shown to produce nonsense when asked with certain prompts, triggering what is referred to as “model collapse”.

1

Not all digital transformations are fully integrated into existing processes. Often, gaps remain where old methods and new technologies don't quite align. What strategies do you employ to ensure that #digitaltransformation is fully integrated into your business processes? #DigitalTransformation #ProcessIntegration #TechAdoption #ContinuousImprovement #InnovationBlend

1