Jumping on to discuss GenAi for the next 45mins. Come join me.

A1. GenAI is an incredible technological advancement. As with any advancement, there are great benefits as well as concerns to watch out for. GenAI will produce a more productive future but we need to be thoughtful on how it’s applied to ensure the best outcomes. #eWeekchat

A2. While interest is high, a cogent strategy for generative AI adoption remains a work in progress for many. We're seeing a lot of effort around improving internal processes and workflows. A major focus for our clients has been on making use of unstructured data #weekchat

A3. 1/2 There’s no question that privacy is top of mind for a lot of companies we’re working with. Samsung’s data leak via ChatGPT comes to mind: bloomberg.com/news/articles/… #eWeekChat

A3. 2/2 There are plenty of ethical concerns to consider around AI-generated content including train source data ownership, potential job displacement, deepfakes and issues related to misinformation demand our vigilant attention. #eweekchat

A4. The future of GenAI will be many smaller, open-source models, fine-tuned on proprietary data with lots of control over what they’re being trained on. This will protect against security and copyright issues and allow companies to control how information is accessed. #eweekchat

Q5. Near term: Targeting LLMs on unstructured data drives the biggest economic value for businesses. Midterm: LLMs agents are going to make huge improvements in automation. Longterm: Major productivity gains with high accuracy task and content generation #eweekchat

Great point. 100% agree. There's so much activity in this space. See Databricks acquisition of .

Such a good point. Samsung banned the use because of this leak: cybernews.com/security/chatg… #eweekchat

A7. Control will be king. The protection of proprietary data is going to be a huge challenge to manage with GenAI. The public models will not be able to support companies' privacy needs. We also need more transparency into what models are being trained on. #eweekchat

An important concern. This article outlines some positives too but there will be a diminishment on accuracy as synthetic data is overleveraged: forbes.com/sites/robtoews… #eweekchat

A8. Good governance and policy development is key. As for implementation, focus on problems that have high business value and lower risks, like extracting value from unstructured data for internal use. Avoid public open models (ChatGPT) with propriety data. #eweekchat

James Maguire A9. LLMs aren’t a panacea; don’t forget about ‘regular’ ML techniques! #eweekchat

A8. The current broadly trained LLMs are only reaching ~90% accuracy on SOME tasks. Getting that last 10% takes a large amount of effort. Be specific on goals and measurements. Also, be patient with iterative development. The effort will be worth it. #eweekchat

A9. The advancements of GenAI have reduced the cost of leveraging and applying AI. This means it's able to be applied to problems that, before, were cost-prohibitive. Experimenting how your business can leverage will yield big rewards as the models continue to evolve. #eweekchat

That time if fell out of a tree, intentionally. A fun edit from my tree BASE jump in Tonsai. youtube.com/watch?v=pKdO9A… #basejump #Thailand

Very excited about Kite Your coding co-pilot kite.com

npm install obscure-lib-only-I’m-familiar-with git commit -am “adding job security”

How to interview a Developer! medium.com/m/callback/twi… medium.com/m/callback/twi…

Awesome!!! My favorite: document.designMode = 'on hackernoon.com/what-you-dont-…