⛔Stop using #data #genai #llms for automation. This is n🚫t only lazy, but it's wrong.
💫It requires a collective effort from all professionals in #ai #data #analytics #optimization #decisions #businessrules to change the way many think about #data and #ai. I refuse to believe that's the best we can do in this industry.
🤣Let's have a look at some stats:
🔸The AI industry spent $50 billion on the Nvidia chips used to train advanced AI models last year but brought in only $3 billion in revenue. [venture-capital firm Sequoia & WSJ]
🔸On average, between 60% and 73% of all data within an enterprise goes unused for analytics. [Forrester]
🔸53% of AI projects make it to production. [Gartner]
What do these tell you? #aigen #llms #data is at the peak of the hype, and this bubble is going to burst. It is going to be worse than the dot-com bubble. Everything around #data #ai #llms #genai and the money spent and invested looks like a Ponzi scheme. Let's have a look at what has happened:
🔸OpenAI released the first Generative Pre-trained Transformer (GPT) to the mass population
🔸It grabbed the public attention and built massive traction
🔸Giant companies with at least around $100M free cash started building more models based on transformers
🔸Every day (almost!), new models are released with slightly better (or worse) scores
🔸Investors pour money on this #aigen technology because it is hot (and is over-hyped)
💣Who are the winners in this game? NVIDIA and the Cloud providers! Of course, they are going to keep and drag the hype as long as possible with all its issues. No one has solved the fundamental challenge of the #LLMs techniques (hallucination). Because this is not a bug or defect, this is the fundamental property of how these models work. Then, to address that, we (as #AI professionals) introduced Retrieval-Augmented Generation (RAG), which is not scalable; it degrades and still fails at a high rate.
😵💫What I am trying to say is to make sure you won't be caught up in this #ai #aigen #llms storm. If you want to use #ai in your organization, start using #decisionintelligence.
✨#ai is a broad field with many techniques such as #optimization #rules #heuristics #machinelearning #graphs that can be safely utilized to solve very hard (borderline impossible) business problems.
🚀You just need to apply the Decision-Centric Approach® which at a high level is to:
1️⃣ Create a holistic business decision model for the business objectives
2️⃣ Decompose the holistic model into smaller and more understandable decision units
3️⃣ Establish the relationships and dependencies between the units
4️⃣ Apply different techniques that are relevant to each unit based on the nature of the problem encapsulated in that unit
✌️Stay safe, stay out of the #aigen #llms hype; you will save money and time!
#decisioncentric
Posted here.
Published April 4th, 2024 at 07:30 am
