AI Agents: Top Five Trends

🗞️ The Tech Issue | December 20, 2023

Author

Qamar Zia
December 20, 2023

☕️ Greetings, and welcome to my daily dive into the Generative AI landscape.

My goal is to streamline this newsletter size, making it a concise, under-five-minute read containing 1500 words or less. Today’s newsletter is 1830 words long which I think is still too long. For those interested in deeper dives, I’ll provide references for extended reading. Most of this content springs from my ongoing research and development projects at INVENEW.

In today’s issue:

  • AI Agents: Top Five Trends

  • Visualizing AI and Tech Hype Using Google Trends

  • How to get a shareable link to a ChatGPT conversation

  • Bizarre Theory Claims ChatGPT Is Suffering From Seasonal Depression

  • Research: Low-Latency ML Inference by Grouping Correlated Data Objects and Computation. (arXiv:2312.11488v1

  • And more

🔔 Please forward this newsletter to your friends and team members and invite them to join. This will help me grow my reach. Thanks, Qamar.

📈 TRENDS

AI Agents: Top Five Trends

  1. Advanced Language Capabilities: AI agents now exhibit remarkable proficiency in understanding and generating human-like language, enhancing applications in customer support, content creation, and more.

  2. Autonomous Decision-Making: These agents are adept at making independent decisions in complex environments, significantly impacting sectors like logistics, finance, and healthcare.

  3. Personalization and Context Awareness: AI agents offer highly personalized experiences by understanding individual user preferences and contextual information, revolutionizing sectors like e-commerce, education, and wellness.

  4. Ethical AI and Transparency: There is a heightened focus on ethical AI practices, emphasizing transparency, privacy, and bias mitigation, aiming to build trust and fairness in AI systems.

  5. Human-AI Collaboration: AI agents are increasingly designed for seamless collaboration with humans, notably in manufacturing, creative fields, and other interactive domains.

🗞️ IN THE NEWS

Google Brain co-founder tests AI doomsday threat by trying to get ChatGPT to kill everyone: Following his participation in the United States Senate's Insight Forum on Artificial Intelligence to discuss "risk, alignment, and guarding against doomsday scenarios," Ng writes in a newsletter that he remains concerned that regulators may stifle innovation and open-source development in the name of AI safety. Read Entire Article Read more at www.techspot.com.

Standigm named as a Tech Innovator in Generative AI in Drug Discovery: Standigm, a company using artificial intelligence (AI) technology for drug discovery and development, today announced that it has been included in a recent Gartner report, entitled, Emerging Tech: Tech Innovators for Generative AI NovelDrug Discovery. Standigm’s technology Standigm offers three complementary platforms for drug discovery and repurposing. Read more at ai-techpark.com.

Kinara Eyes GenAI With 6-W Edge Chip: AI chip startup Kinara can run generative AI models with up to 30 billion parameters on its chip for edge servers, laptops, and game consoles The post Kinara Eyes GenAI With 6-W Edge Chip appeared first on EE Times. Read more at www.eetimes.com

Microsoft’s Stuffing Talking Generative AI Into Your Car: Remember those talking GPS monitors from the early-to-mid-2000s? They may soon be back with the help of Microsoft's artificial intelligence capabilities, even though it's unclear who asked for them. As Reuters reports, the GPS company TomTom is teaming up with Microsoft to equip cars with a generative AI assistant that will, and we quote, allow drivers […] Read more at futurism.com.

Bizarre Theory Claims ChatGPT Is Suffering From Seasonal Depression: The days are getting shorter, temperatures are plummeting, and for many, the seasonal depression associated with a long and arduous winter is really starting to settle in. And it may not just be humans struggling with the lack of vitamin D. OpenAI's popular ChatGPT chatbot may be suffering from the same fate — that is, […] Read more at futurism.com.

The promise and peril of AI at work in 2024, according to Deloitte's Tech Trends report: As the biggest year for Generative AI development comes to an end, we use those learnings to take a look at what to expect during the next 12 months. Read more at www.zdnet.com.

⚙️ ENTERPRISE

Choosing a GenAI partner: Trust, but verify: Enterprise executives, still enthralled by the possibilities of generative artificial intelligence (GenAI), more often than not are insisting that their IT departments figure out how to make the technology work. Let’s set aside the usual concerns about GenAI, such as the hallucinations and other errors that make it essential to check every single line it generates (and obliterate any hoped-for efficiency boosts). Or that data leakage is inevitable and will be next to impossible to detect until it is too late. (OWASP has put together an impressive list of the biggest IT threats from GenAI and LLMs in general.) Read more at www.computerworld.com.

Four priorities to set for generative AI in the year ahead: A recent ExtraHop survey of 1,250 IT and security leaders highlights key concerns and strategies for managing generative AI risks. The main findings include the need to understand AI-related risks, the ineffectiveness of outright bans, a call for government and industry standards, and the importance of basic cyber hygiene and cross-functional collaboration. These insights are crucial as generative AI reshapes technology and security landscapes. Read more at SCMagazine.

🏃 QUESTIONS

These six questions will dictate the future of Generative AI:
In 2023, generative AI, reminiscent of the early internet, is reshaping how we interact with technology. It's enhancing tasks like report writing, triggering debates on its ethical implications, and raising critical questions about bias, copyright, job transformation, misinformation, and societal impacts. This tech is evolving rapidly, with immense potential and challenges, echoing the transformative and disruptive nature of the dot-com era. Read more at www.technologyreview.com.

🏢 TIPS

How to get a shareable link to a ChatGPT conversation: ChatGPT conversations can be lengthy and challenging to share. Fortunately, you can share a link to a conversation, accessible without a ChatGPT account. However, the share option only appears for recent sessions and shares the entire session's content. Anything added after creating the share link remains private. This feature is useful for focused sessions you plan to share. Read more at www.zdnet.com.

Visualizing AI and Tech Hype Using Google Trends: A tutorial on how to create slope-graph visualizations for assessing technological trend shifts, such as virtual reality and generative AI. Read more at towardsdatascience.com.

📦 USE CASES

In 2023, Rubric, creators of GitHub Wrapped, launched "Year in Code", a unique, AI-generated video project for GitHub users. Leveraging LangChain and OpenAI GPT-4-turbo, the team produced personalized videos, reflecting users' GitHub activities. This ambitious project utilized GitHub's GraphQL API, Remotion for video creation, AWS services, Three.js, and Next.js, showcasing cutting-edge technology and innovative use of AI in celebrating and inspiring the developer community.

📚LEARNING

Benchmarking RAG on tables: Key linksLangChain public benchmark evaluation notebooks:Long context LLMs hereChunk size tuning hereMulti vector with ensemble here MotivationRetrieval augmented generation (RAG) is one of the most important concepts in LLM app development. Documents of many types can be passed into the context window Read more at blog.langchain.dev.

🔬 RESEARCH

Low-Latency ML Inference by Grouping Correlated Data Objects and Computation. (arXiv:2312.11488v1 [cs.DC]): ML inference workflows often require low latency and high throughput, yet we lack good options for addressing this need. Techniques that reduce latency in other streaming settings (such as caching and optimization-driven scheduling) are of limited value because ML data dependencies are often very large and can change dramatically depending on the triggering event. In this work, we propose a novel correlation grouping mechanism that makes it easier for developers to express application-specific d...
Read more at arxiv.org. Published on 20-Dec-2023

Reference:

💡 REVIEW

In the concluding installment of their 'Generative AI Openness' series, the authors delve into the openness of Large Language Models (LLMs). They juxtapose the more restricted approaches of giants like OpenAI and Google with the comparatively open stance of Meta. This exploration covers a range of LLMs, scrutinizing the extent of openness in their development, collaboration, and accessibility. The article sheds light on the varied levels of accessibility across models such as Alpaca, Vicuna, GPT-J, Falcon, and BLOOM, emphasizing the challenges in reproducing these AI systems due to diverse restrictions on data and code access.

Reference: DEV.TO How Open is Generative AI?

📣 DATA
In search of the missing piece of generative AI: Unstructured data: The spotlight in recent years has shifted towards unstructured data (text, graphics, documents, IoT streams), which hold immense, untapped value. The database industry significantly adapted to surface these assets. Despite 90% of enterprise data being unstructured, only 46% of organizations actively extract its value. The emergence of generative AI further emphasizes the importance of unstructured data. Companies proficient in managing unstructured data are well-positioned to leverage generative AI for deeper insights. This trend is reshaping data management strategies, focusing on unifying data platforms for analytics and AI, essential for harnessing generative AI's potential.

Your Feedback

I want this newsletter to be valuable to you so if there's anything on your mind—praises, critiques, or just a hello—please drop me a note. You can hit reply or shoot me a message directly at my email address: [email protected].

Join my community by subscribing to my newsletter below:

Reply

or to participate.