Sourcing Journal’s Post

Chief Revenue Officer Pedro Ramos met with Meghan Hall at Sourcing Journal to discuss our newest solution–Incident + ORC Intelligence! Powered by generative AI, this tool helps identify #organizedretailcrime (ORC) rings and repeat offenders, offering retailers a proactive approach to tackling one of the industry's biggest challenges. Check out the article below! 👇

Interesting news coming from Appriss Retail! Such tools will be popping up, as this scheme is not only useful for detecting connections between different organized retail crime groups, but many others. The key is how the data gets fed in. #frauddetection #ai #chatgpt

A little video explaining a new way generative AI is being used to tackle organized retail crime enforcement! Check it out, and read the story:

ORC Intelligence helps you capture #OrganizedRetailCrime incidents, match related cases via Generative AI-powered technology, and collaborate with other investigators to maximize efforts, and shut down #ORC groups faster, saving you time and additional losses.

If you feel a bit dwarfed by AI right now, if you’re unsure about whether an LLM that sounds very much like it is ‘thinking’ is actually thinking, and/ or if you’re increasingly relying on the plethora of emerging AI-based tools to do your job, have a look at this article, and try trusting it to do heavy lifting less often. 😉 Some things to reflect on (which is something a machine can’t do…): 1. Machines execute code as designed. They exist in a quasi-deterministic reality, bounded by an existing body of information (‘knowledge’), with no perception of space or time. 🗿 2. Machines do not connect any meaning from one moment to the next, have no way to identify with anything, and only ‘learn’ according to strict parameters outside their control. 🎲 3. Machines are like magic goo, through which obscene quantities of human language pass, leaving traces of meaning to be gleaned and reconstructed, so that they only ever combine the old in different ways, and when they are judged by the observer to have produced symbols that mimic coherent speech, they are directed to continue doing so. 📝 AI isn’t mysterious, and it’s not achieving consciousness. Letting it do your homework too often isn’t smart…it’s just short-sighted and lazy. 🧮<🧠 Write your own code. Add real value.

BREAKING: OpenAI’s big reveal is here & it’s indeed a voice interface across media! Check out the demo here 👇 ☑️Text to speech is indeed what many of us were expecting to announced today. But is it impressive enough? What you see here compared to the voice interface on ChatGPT before: ⚡️Lightning fast speech-to-text-GPT-speech processing! None of the terrible lag times that made it previously useless. 🦹♀️ Emotions & drama baby! Hear the laughter, feel the smile? You can sway to the sing-song or robotic twang. A taste of Emotional AI is here — from sensing & delivering the emotional experience on demand. 🌠 A new portal to experience & interact with GPT: voice! This is what we were expecting for a while. No more typing All this is great — not mindblowing but great — but what is the real verdict? ➡️TLDR: If it works consistently & realiably well across the billions of users, the voice interface is definitely useful! Only time will tell if this is a overnight gimmicky feature we will tire of or something that can impressively upgrade how we interact with Gen AI Me, I’m hoping for the latter! ✨

ChatGPT-4o has offered faster processing speeds, but it has also shown signs of unchecked racism and environmental harm. Adah Parris (she/her) explores how incorporating traditional knowledge into the development of AI can make it more just. https://lnkd.in/ewuAw__y

In the deluge of news and hype about Generative AI, it can feel overwhelming. But I have one simple suggestion to up your personal GAI game: “First, ChatGPT” No, I don’t mean simply try it out, as my friend Jay Leib likes to joke- there isn’t anyone under 85 years or over 85 weeks that has not tried it… Rather… shift your default. Unpavlov yourself from simply Googling and make GAI your first resort. For me 25 years of Googling was fully supplanted as a default in just 25 weeks with the GAI renaissance. Why do this? To refine your prompt engineering capability on low risk tasks. Need a recipe, want to know an actor in a movie, want to up your brisket game? First, ChatGPT. Treat this GAI first approach as a training ground to level up your ability to get GAI to do what you want. By shifting your behavior patterns you increase your opportunity to learn iteratively how to effective prompt. Like riding a bike or learning a new language, it is all about trial and error and itteration! Read more GAI takeaways from International Legal Technology Association (ILTA) #Evolve in this great Legaltech News article: https://lnkd.in/dn-r-kKe #ai #legaltech #weareilta #technocat

Imagine walking into a restaurant where one chef does everything: greets you, takes your order, cooks the dish, serves it, and then clears your table. Seems insane but that's exactly what is happening when asking ChatGPT to solve a problem end-to-end. A more natural flow is to have a number of LLMs running together, validating and verifying each other. Typically this is known as an Agentic workflow. Agents are architectures that use multiple AI Models, tools and processes to extract the correct information. A great example is when given a coding problem OpenAIs GPT3.5 scored 48.1%, and their state-of-the-art GPT4 scored 67%. Taking the same coding challenge into a Agentic flow using multiple GPT3.5 models, that model was able to score 95.1% Agentic flows are able to plan, test and evaluate results before producing results. Sort of like having a team of people running a restaurant. See the short video on agentic workflows from Andrew Ng. https://lnkd.in/gZtvp3qK

Let's talk data. We hear a lot about ChatGPT and other tools, but analysis shows another 62 platforms with over 1 million monthly hits. Combine this with search data and we get an insight into 1) how big Generative AI already is and 2) how diverse its applications. Now, BRB, I'm going to learn more about Brainly :) and how it could influence a happier future of work. Full report: https://lnkd.in/dfwV3Vx7 Antony Slumbers Shane Dillon Rebecca Hinds, PhD Robert Scoble