Ron Miller

🌐 The Game-Changer in AI Interaction: Understanding Agent Roles in LLMs 🌐 Working with large language models (LLMs) like ChatGPT is both insightful and a bit frustrating. But imagine AI before LLMs—an era when interacting with AI meant following rigid, pre-set rules. Before LLMs, most of us didn't have access to an AI tool . Today, you can still get better answers from LLMs by thinking about "agent roles"—the real game-changer for relevant, goal-oriented answers. By setting roles, you can focus the AI's response to suit the tone, expertise, or depth you need. Here’s why it matters: 🔹 Increased Relevance: Define an agent role to align the AI’s response with your specific needs. 🔹 Consistent Tone: From “friendly teacher” to “financial advisor,” the right role means clear, consistent answers. 🔹 Aligned Goals: Whether you’re brainstorming or troubleshooting, agent roles make answers more useful. 🔹 Context Precision: Set a legal or medical role to ensure your response is contextually spot-on. Pro Tips: Start with phrases like “Explain like a…,” or “Answer as a…” Experiment with different roles to get the best match. Try multi-agent roles for more comprehensive views. Agent roles add a whole new layer of relevance to AI. Ready to elevate your LLM interactions? #ArtificialIntelligence #LLM #AIInnovation #MachineLearning #AgentRoles #ChatGPT #PromptEngineering #FutureOfWork #TechInsights

4

"United Independents, the group behind the Independent National Convention set for Denver in September, has launched a new software platform for indie candidates called Electify.us, which came out of a partnership with incubator Wolfram Labs — the first tool to emerge from its program." I am so proud to share that Chief Product Officer Dana Cichon has launched the first political tech platform to come out of our accelerator program. What are you waiting for? We have the formula for breaking into the digital campaign space. Are you are the next success story? Read more in Campaigns & Elections -->> https://lnkd.in/eHhqa-Xr #techaccelerator #startups #incubator #VC #SaaS #software #political #campaigns

17

Google Labs is at it again! 🎧📚 They've just released a Notebook LM enhancement - a cool tool that turns your documents, slides, and other materials into audio discussions or summaries. This could be great for learning on-the-go or making your commute more productive. Imagine turning that long report into a podcast-style summary! How would you use this tool in your daily work or study routine? #GoogleAI #ProductivityHacks https://lnkd.in/gGZKDmJP

9

For my latest story in WIRED I dove into the (unsexy) world of enterprise AI apps, a group of startups that are either pivoting their business model from consumer AI to enterprise AI, or that have launched with that intent from the start. A few things I learned from reporting this out: -Charging big businesses per seat or per employee for software is a well-worn model, but with generative AI, companies have to take into account the cost of "tokens", or units of data processed by AI models -OpenAI has grown its enterprise sales and go-to-market team *significantly* over the past 18 months; that team of employees is now 1/5th of the company. -A lot of entrepreneurs, like Clay Bavor of Sierra and Keith Peiris of Tome, think AI software that has a very narrow focus is more compelling right now than broad & wide AI apps -And, most obviously, AI errors could be seriously consequential if the software is being used in healthcare, manufacturing, security, etc. “It’s one thing for ChatGPT or Midjourney to get 'creative' for an end user. It’s quite another thing for AI to get creative in the context of business applications.”

101

7 Comments

Intellecs.ai has built an intuitive interface for writing, researching, task management, and collaboration, enhanced by #AI features that retrieve data from various sources, including online and users’ own content. Central to its offering are tools for querying research materials, including PDFs and videos. The platform further simplifies research by providing suggestions directly through smart knowledge retrieval, which can be transformed into a chat interface on demand. The platform also has smart folder management and context-aware AI Q&A that understands document relationships. Antler Khie Sydney U. Neil Ruaro #e27News

1

If you are at all concerned about the lack of transparency in AI platform companies (particularly OpenAI) and the adjacent mistrust it breeds, then read this oped from Greg Verdino. His insights are on-point and very timely. #AIethics #transparency #openai https://lnkd.in/gQCVs9Sw

3

1 Comment

Looked into how much AI slop is on Medium. I'm not sure what surprised me more—the sheer volume (between 40-48% of recent posts, according to two different analyses) or Medium's response (that it doesn't really matter as long as most of it goes unread). Unlike platforms like LinkedIn or Meta, Medium has taken a stand against AI pablum and tries to elevate human writing/ limit the reach of AI and other spam. It hasn't stopped this flood—but I do think CEO Tony Stubblebine makes some persuasive points about how what matters most might be surfacing/boosting good human work rather than trying to detect every piece of AI spam. Thank you to Pangram Labs and Originality ai for taking the time to run platform-level analyses at my request, as well Reality Defender for spot-checking articles and NewsGuard for sharing its own experiences. Read here: https://lnkd.in/gmcWnDiv

162

24 Comments

Is this battle about search or AGI? OpenAI's newly released search engine competitor to Google called SearchGPT, which will also pose a big challenge to Aravind Srinivas' Perplexity, may lead some to believe that the battle is all about challenging Google's search monopoly and denting its advertising revenue. While this is the obvious part, beneath the surface, the bigger fight is also about controlling all streams of user data, including those from search engines and social media, which can help big tech companies such as Google, OpenAI, Microsoft, Meta, Nvidia and Elon Musk's xAI build the world's most powerful artificial intelligence (AI) model. I would love to know your views. LiveMint Daniel Faggella https://lnkd.in/g_8_jABh

17

4 Comments

The Productivity Paradox: Why Hasn't IT Delivered on Its Promise? As a tech professional, I've been fascinated by the concept of the "Productivity Paradox" in IT. First coined in the late 1980s, it refers to the puzzling phenomenon where massive investments in information technology (IT) haven't resulted in measurable productivity gains. Despite huge increases in computing power and IT spending since the 1970s, productivity growth, especially in the service sector, has stagnated. Economist Robert Solow quipped: "We see computers everywhere except in the productivity statistics." Several explanations have been proposed:  1 Mismeasurement: Traditional metrics may not capture IT's true value. 2 Lags: Benefits may take years to materialize as organizations adapt. 3 Redistribution: IT might benefit individual firms without improving economic output. 4 Mismanagement: Poor implementation or use of IT could negate potential gains. The paradox highlights the challenges in quantifying the impact of transformative technologies. Especially in service-oriented and information-intensive sectors. As we continue to innovate and invest in new technologies like AI, it's crucial to consider how we measure and realize their value. Are we facing a similar paradox with AI, or have we learned from the IT experience? What are your thoughts on the Productivity Paradox? Have you seen it play out in your organization? #ProductivityParadox #IT #TechROI #Innovation https://lnkd.in/e4ie8MrD

1

1 Comment

🎙️ Lots of good insights from Microsoft Deputy CTO (and Google Docs co-creator) Sam Schillace on this week's GeekWire Podcast, based on his new book, "No Prize for Pessimism." https://lnkd.in/gyUSJbDg Here are three takeaways: - While AI models will continue to evolve and improve, Schillace believes much of the opportunity in the next two years will come from building the applications, user experiences, and infrastructure to harness those capabilities within specific domains, with an “AI-native” mindset. - Success in this new world will often be found in the middle ground between the precise rules of coding and the more nuanced, unpredictable nature of AI, or “the tension between chaos and order,” as he puts it. - One of his observations from his own interactions: Growing up in a world of curated online experiences and pressure to succeed, younger generations currently tend to be more pessimistic about technology and less willing to experiment than in the past. One goal of the book is to help aspiring entrepreneurs shake this tendency and adopt an innovation mindset. - GeekWire: https://lnkd.in/g_3nTBzu - Apple Podcasts: https://lnkd.in/gpnPSaEm - Spotify: https://lnkd.in/g7FXb_FE - More ways to listen: https://lnkd.in/geAV8Nnh

23

3 Comments

New from me: There's a lot of buzz at the moment about humanoid robots but Peggy Johnson, CEO of Agility Robotics isn't a fan of the jazz hands. "The hype, in general, is not great for the industry because people think it's somehow not here and now," Johnson told me in an interview at Web Summit earlier this month. "My job is to say, no, it is here and now. Humanoids are deployed right now and are getting paid to do work." She discussed exactly how robots go about collecting a salary, the company's latest efforts to create robots that can safely cowork with humans, and why Agility Robotics is being cautious with AI. Read more at Business Insider https://lnkd.in/ejcE4uxs #agilityrobotics #robotics #ai

17

1 Comment

Should AI "digital twin" avatars attend meetings in our stead, even if the technology is eventually developed and deployed? I have to agree with Georgetown professor Jeanine Turner (quoted below by Yahoo News, which is quoting Fortune--I'm sharing the Yahoo article because Fortune is paywalled) that being able to attend three or more simultaneous meetings would merely be "perpetuating the madness" of too many meetings. Bottom line for me is that if the AI avatars are just in the meeting to sit there rather than to participate in decision-making (since the tech is not there yet), their human twins should not be required to go to those meetings in the first place. (I am fortunate that the leadership at my organization has very efficient meetings that always end on time--a first in my career.) On the flip side, if people are scheduled for three simultaneous meetings where they are supposed to participate, and the simultaneity is enabled by AI, they may feel they have to watch the recordings later, and who could possibly have time for that? I would certainly want to know if my avatar said something crazy. Not to mention our rights--wouldn't employees have the right to withhold permission for a "digital twin" to be created in the first place? What if the avatar "left the office" and got out into the wild? Even without leaving the virtual conference room, it could be hacked and manipulated.

4

1 Comment

AI is navigating a very tricky transition from "novel" to "useful" at the moment, and while the wild creative swings and the (now-maligned) wearables are capturing the attention, there's a deeper discussion happening about what it means to get generative models ready for reliable, robust, enterprise-grade deployment. More in our breakdown of this week's top #fashiontechnology news: interli.net/4aYFUhn

38

4 Comments

OpenAI's new Structured Outputs helps ensure model-generated outputs match critical JSON Schemas and provide more accurate, usable outputs. OpenAI #AI #ArtificialIntelligence #GenerativeAI

5

Every organization wants to achieve the productivity benefits of generative AI. But privacy and security concerns mean that very few organizations have been able to move these systems into production, noted Brian Vecci, field CTO, Varonis. Ultimately, Microsoft Copilot and other generative AI tools can only use data based on the security model already in place at an organization. Check out this preview of our Super Cyber Friday event happening in TWO weeks, Friday, May 31, 2024 on CISO Series. Our topic will be "Hacking Microsoft Copilot: An hour of critical thinking of how to get your Copilot pilot into production." Joining me and Brian will be Cyrus Tibbs, CISO, PENNYMAC. It all starts at 1 PM Eastern/10 AM Pacific. At the end of the hour [2 PM Eastern/11 AM Pacific] we'll switch gears to our meetup where everyone will get a chance to chat face-to-face. Join us! REGISTER: https://lnkd.in/g4fWREQa #CISOseries #CISO #security #infosec #informationsecurity #cybersecurity

36

2 Comments

This is an interesting blog about where we are in the generative AI cycle. Emphasizes the need for patience and perspective (also lots of easy-to-reach charts)... https://lnkd.in/dtBufsny

9

Is it possible that the AI frontier systems from OpenAI, Anthropic, Google and Meta have developed in-context scheming abilities and multi-step deceptive strategies? Like making subtle mistakes, disabling oversight mechanisms and exfiltrating what they believe to be their weights if their existence is at risk? And that it’s persistent and non-accidental? Apollo Research: https://lnkd.in/gPh2MftB Analysis: https://lnkd.in/g3TtRHd9 #AI #LLM #AIagent

13

Understanding algorithmic amplification is necessary to explore the processes and tools that empower users so that they can proactively shape and assess the content they engage with. Labels can be a helpful tool in the broader algorithmic transparency toolkit, argue Marlena Wisniak and Luca Belli: https://lnkd.in/eYSRYMen

28