"Age of Copilot"​ + "Age of Responsible AI"​ + "Age of Vertical Disruptions"

Top 10 Generative AI Highlights and Insights 4.2.2023

I am on the road now, but I still cannot help reflecting on last week's generative AI news.

1) Set the right expectation: the current "Age of AI" is an "Age of Copilot."

I have not done any weekend parties for at least three years, and my very first one after a long pause was to attend a party organized by a Stanford University Graduate School of Business alumni friend. No surprise: It was all about generative AI. Every guest gets to write down three jobs that are going to be replaced by #AI and three that are the most secure. I explained at the party that #AI is our copilot for a long time to come. It means we humans need to know "how to fly", and we should leverage AI for what they are good at as much as possible, but there will be "humans-in-the-loop" for quite the foreseeable future at least.

It occurred to me that the #GenerativeAI industry needs to set the right expectations: the "Age of AI" is the "Age of Copilot," and it is not the "Age of Autonomy" yet.

2) Microsoft introduced Security Copilot: security industry professionals get to tighten their seatbelts!

Microsoft introduced an AI assistant tool for security analysts to help them triage data quicker and search for potential threats or vulnerabilities faster, among many other tasks. 

The marriage between #GPT4 and Microsoft's security model can answer queries from security analysts such as: "How can I improve my security posture"; "What are the trending threats"; "Which alerts are being triggered the most"; and "Tell me about my latest incidents".

I don't think the current demo is super impressive just yet, but I definitely see a seismic shift in the #Cybersecurity industry, so tighten your seatbelt now!

3) "To pause, or not to pause" -- "Age of AI" must be accompanied by the "Age of Responsible AI"

A group of high-profile tech leaders, including Tesla's CEO, #ElonMusk, called for a six-month moratorium on the development of advanced #AI #LLM above #GPT5. Was this excessive caution or the sounding of a much-needed alarm?

I joined a couple of #CIOs and experts from Microsoft, Swisscom, and HiSolutions AG to discuss some of the concerns behind advanced AI. Read my takeaways here on CXO REvolutionaries.

I also want to add that, according to a friend who talked to the author of the moratorium, this is not likely going to happen: "There's no way Microsoft is letting this momentum go.” They have waited years for this and have invested money in the right places. If you ask me, I don't bet on any pause by OpenAI either.

4) Goldman Sachs just released their research: as many as 300 million jobs could be affected by AI.

In the United States and Europe, approximately two-thirds of current jobs “are exposed to some degree of AI automation,” and up to a quarter of all work could be done completely by AI, the bank estimates.

As you can see, the following job sectors are at the greatest risk:

-Administrative: 46% at risk

-Legal: 44% at risk

-Architecture and Engineering: 37% risk

5) Google's #bard used #ChatGPT to train the model, according to a former employee. Is it embarrassing for Google to do it, or is it the future of data labeling?

Of course, Google denied the allegation. I personally don't care as much, even if it turns out to be true. I am happy to see each of us stand on the shoulders of a giant so that we can move the world forward faster. In the chatbot landscape, Google is the underdog today.

6) Another industry that #GPT is disrupting is the healthcare industry.

Here is a new book on the "Benefits, Limits, and Risks of GPT-4 as an AI Chatbot for Medicine."

7) Bloomberg introduced #BloombergGPT, a large language model for the financial industry.

It is another profound milestone as it could imply the boom of vertical models; however, the jury is still out on whether all verticals need their respective LLMs.

Bloomberg claimed that the BloombergGPT model outperformed similarly-sized open models on financial NLP tasks by significant margins, without sacrificing performance on general LLM benchmarks.

8) I got a taste of the new OpenAI #GPT4 plug-in for the first time. It didn’t go as well, but it is only the first week after all!

The Expedia Group plug-in didn't quite work. I asked for hotel availability in one zip code in Miami, and it told me no room was available during a time when the Expedia website told me rooms were still available.

9) An open source race towards replicating ChatGPT: Colossal-AI joins the party

Colossal-AI not only introduced ColossalChat but also the #RLHF #framework. I don't think it implies the world can replicate a model as capable as #ChatGPT yet. However, plenty of parties are working towards it, and there is a possibility to see the magic happen in the open-source wonderland in 2023.

10. Task-driven Bots are likely going to proliferate.

Here is just one of many examples of how task-driven autonomous agents are being implemented. It is an article by Pinecone, but the particular choice of vector database does not really matter.

Published by