Observations from early generative AI engagements

Much has already been written about chatGPT and the potential impact of AI on business. Here are some observations from leading engagements across Canada on generative AI:

  1. Tremendous excitement and engagement: ChatGPT inspired interest. However, it can be argued the community really started when Meta launched Llama. Less than 2 weeks later, Llama was leaked to the public, accelerating the pace of innovation. Shortly thereafter, several camelid family-name models followed. Stanford released Alpaca on March 13, followed by Vicuna, and Dromedary. Enterprise interest is palpable.
  2. Cautious enterprise experimentation: In the last 3 months, I have worked with about a dozen clients with wide variety of implementations. Prompt engineering has entered general vocabulary, and Enterprise Architects become familiar with RAG and Vector Databases. Many organizations have issued policies banning use of generative AI. Italy banned and then allowed access to ChatGPT. I have seen Canadian clients embark on learning, grasping the nuances of RLHF, and experimenting with low-hanging fruit use cases such as customer care, customer segmentation, hyper-personalization, and more interesting use-cases such as AI-assisted metadata generation.
  3. Debate rages on size: Clients are enamored with model parameter sizes. A whole new technical jargon is now available with GPT-3 at 175B parameters, Palm-2 with 540B, Vicuna with 13B, Llama with 7B, 13B, 33B, and 65B parameters, etc. Bigger is not always better though. Similar there are discussions about context size and perceived drawbacks. Summarization techniques and sophisticated chained prompts, maintaining vector databases to keep embeddings for custom documents and then “searching” across them by some similarity metric, fine-tuning the LLM with custom data when possible or developing custom smaller LLMs for this particular data (again, not an obvious task) are some of the techniques employed so far.
  4. Implementation architectures are emerging: Reference architectures are emerge for Generative AI platform, and applications. However, early implementations are still grappling with banal questions about the location where the model is hosted for inferencing (with the model provider, or behind the enterprise firewall?), data sharing, use of vector databases, etc.
  5. Concerns about HAP & Ethics: Much has been written about the dangers of stochastic parrots already. Many organizations including UNESCO have released AI Ethics frameworks. Concerns here remain about the underlying biases in the training data on gender, race, religion. Organizations in particular are also worried about Hate, Profanity and Abuse. More to come on this topic.



Richard S.

Financial Services Tech Executive & Angel Investor

1 年

Nice share Manav Gupta .tks

回复
Imad Lodhi

Co-Founder CustodyMate | Co-Founder Jnana Analytics | Global Sales & Delivery Executive | Canadian Delivery Excellence Leader | MEA Services Excellence Leader | CEE Delivery Leader | Contact Center Practice Leader

1 年

Good article. We have to look at the audience. The technical people are all excited and jumping with joy about models, model size, security, governance, ethics, etc. But business leaders are looking from the lens of business - how can this technology help me make more money and maximize profits? Despite all the mentioned observations, the true value of generative AI lies in its applicability to address the specific needs of businesses. And that is where i see details and depth lacking from many on LI.

回复
Arunachalam Jayaraman

Vice President - Multi-Cloud | DevOps, SRE, AIOps, Automation | Agentic AI | Gen-AI, Prompt Engineering | Digital Transformation | Cloud Security | Cloud Reliability & Sustainability Engineering, Consulting

1 年

Thanks, Manav, for sharing your insights and observation, I couldn't agree more with you, on "Bigger is?not always better though.?" when it comes to the parameters, :-) Unintentional bias in the data and intentional bias forcefully included to "makeover" the context and sentiments are too scary. Ethical frameworks for AI usage should be mandated and enforced by the respective governments. Specifically, EU and Singapore are pioneers in regulating some of these. Looking forward to the whole AI governance wihtout compromising the extremely well-equipped technical capabilities of Gen-AI

回复

Thanks Manav for sharing your views points and insights. Bigger is definitely not going to be better when it comes to LLMs. I might want to add that governance will be key as we move forward with applications of FMs to ensure quality outcome. Further multimodal FMs are what will provide value to enterprises in the applications space. Beyond this policy framework to Gen AI will need to be developed by the consumers of GenAI to ensure it takes into consideration societal, ethical, legal aspects that are above and beyond economic calculations.

Jianli Ren

Executive Data Scientist | Data & AI/ML in Hybrid & Multi-?? | ?? + ?? on the ground

1 年

Thanks Manav for sharing your insights and observations. It’s a very good summary and reflection of the what's happening at the #GenAI front.

回复

要查看或添加评论,请登录

Manav Gupta的更多文章

  • Defining a company's culture

    Defining a company's culture

    It's safe to say we all talk about, and understand, the culture of the companies we work for. It's harder however, to…

  • Blockchain Production Deployments

    Blockchain Production Deployments

    All technologies have a hype-cycle curve conceptualized and branded by Gartner. Every few months there is a discussion…

    3 条评论
  • Blockchain Trends for 2019

    Blockchain Trends for 2019

    2018 has been a banner year for blockchain as industries spanning the gamut from retail to shipping embraced the…

  • Emerging Patterns in Blockchain

    Emerging Patterns in Blockchain

    Over the course of 2018, here are my observations in blockchain projects: Decline of the Proof-Of-Concepts: From 2015…

  • Blockchain for the Enterprise

    Blockchain for the Enterprise

    Some people have asked for the table of contents for my book Blockchain for the Enterprise. So here it is! Chapter 1:…

    6 条评论
  • Shapes of an individual

    Shapes of an individual

    I get asked about leadership, mentoring, professional development and general guidance. One common theme I keep coming…

    1 条评论
  • Blockchain and related technologies

    Blockchain and related technologies

    Now that I've been associated with a few blockchain projects for enterprise clients across several industries (FSS…

    1 条评论
  • What makes a good blockchain use case?

    What makes a good blockchain use case?

    Over the last two days, I have delivered blockchain-101 lectures to about 400 people. One of the predominant questions…

    4 条评论
  • Overview of Hyperledger

    Overview of Hyperledger

    One of the most common questions I get asked is 'What is hyperledger?' soon followed by 'Is hyperledger IBM's…

  • Predictions for Cloud Computing - 2017

    Predictions for Cloud Computing - 2017

    Now that 2016 is almost over, I wanted to sit down and transcribe the trends I'm seeing in the field including the…

    2 条评论

社区洞察

其他会员也浏览了