It's time to climb that mountain | You and AI News #5

How to up-level your AI skills at work plus latest news

In Ep #5 “It’s time to climb that mountain” ⛰️

Read time: 8 mins

Lots of content this episode to help get more into AI. We mix news with helpful pointers on how to make AI work for you. What a time to be alive! 🥳

  1. What’s your path from here to AI expert?

    How to get from Level 100 to Level 400.

  2. The new AI workforce?
    Today, GPTs. Tomorrow, Agents, and Future, AGI?

    What you can do today, and get ready for what’s coming next.

  3. How can you lead on AI in the workplace?

    Roots-up initiative is great, but top-down exec engagement is key – here’s how.

What’s your path from here to AI expert?

How to get from Level 100 to Level 400.

Every journey starts from where you are right now. The route and destination are ahead of you and under your control.

No matter what your current level, keeping up with the AI news is important because it moves faster than any tech we’ve seen before. What was black last week can be blue this week and red the week after.

If you want to move from Level 100 Novice to Level 200 Intermediate, then here’s some alternative things to learn from and try.

AI Pitch Decks are not just for investors. 💸

You can learn a lot from AI pitch decks: Who’s pitching? What are the pitching? HOW DOES IT AFFECT ME NOW AND IN THE FUTURE?

Make your job easier with these 7 ChatGPT hacks 🔨

  1. Use AI for research. ChatGPT Plus has web browsing. Search is a fire-and-forget, but ChatGPT is an interactive tool that does things you ask it to. It’s like a research intern.

  2. Write reports and other content. After your research, AI is great at outlining different content types. It can also tell you what to put into the sections. But the human touch is usually essential to turn AI’s output into a quality deliverable.

  3. Transform meeting transcriptions. When you get a meeting transcription, put it through AI and get a summary, key points, and action items.

  4. Ask AI’s opinion on your work. Upload your report PDF, business plan, cash flow statement, and ask AI how you can improve it. Imagine you and AI are sat in front of a whiteboard critiquing your work.

  5. Your coding assistant. Whether you’re a senior engineer looking to save some time, or a beginner engineer trying to get better, tools like Github’s Codepilot boost your productivity.

  6. Ask AI what charts mean. You can upload a picture of a chart and ask ChatGPT what it means. People are using this for stock charts, website performance, and AI can even tell you what the image is.

Other ideas in Business Insider.

Commercial real estate is having an existential crisis, but the boom in artificial intelligence may be providing a bright spot for tech hubs.

New data from a CBRE study first released to Yahoo Finance showed that the tech sector reclaimed the pole position as the top industry for office leasing growth in 2023. Read more.

AI is getting faster — are you?

The new AI workforce

Today, GPTs. Tomorrow, Agents, and Future, AGI?

OpenAI’s GPTs have been so successful they had to pause ChatGPT Plus signups because they can’t count their money fast enough. Ok, that last bit was a joke, really it was because they don’t have the compute capacity to cope with demand. Why are GPTs so popular?

Imagine a GPT as a box you can create and store your AI goodness in.. then share that box with the world. The AI goodness is your instructions, data, and actions that can connect to an API or website. Last week we used Weather Visual Artist as an example.

Now look at your Jobs To Be Done. Imagine creating a GPT for one of those JTBD. Imagine it for all your JTBD. Imagine it being like a new hire, someone working for you to take the strain and help you get more done. Filling in for skills you don’t have. GPTs can 10x you.

 BUT BEFORE YOU GET TOO EXCITED, let’s temper the hype with reality, courtesy of Alex Volkov.

Alex Volkov is the creator of the Visual Weather Artist GPT. Here’s a summary of what he cautioned about the realities of GPTs “today” (because they will evolve quickly):

- Editor is really basic, not even tab works, which if you're writing yaml or json, you kinda need. I keep hitting tab and the focus moves.

- Authentication is not robust enough, I've tried maybe 5 different services so far, all require auth key in URL, and OAI doesn't have a good solution for that. You can add the secret value to the "default" section in spec, but then the UI shows your key to the user.

- Debugging: maybe the no 1 reason I got so so frustrated, it's impossible to debug actions that went wrong. You can try to ask your GPT, "what's wrong" or "recreate the action call with a CURL example" but it should be much easier to understand and fix

- Slow: I don't know why actions are slower than browsing, it sometimes takes up to a minute for the action to show up and ask the user "do you want to send this data"?

- Actions GPTs are greyed out on IOS for now. A user can start a conversation on web and the continue on IOS, but can't start a new one, why? 🤔

- Examples are also basic, I'm hoping they are working on some kind of builder for that maybe?

- The whole OpenAPI spec, it was nice for plugins, but are we really sure that this is the "low code" that the online users need?

- Why only 1 action? Why can't I add multiple? The model can handle it, let me add a few please 🙏

Agent Smith? Is that you?

Are AI agents real?

GPTs can be thought of as a step close to AI agents for the masses. AI agents aren’t a new concept. Anyone at AI Level 300/400 will have spent time with Hugging Face and LangChain (more on that in future episodes, so subscribe to get that).

What are AI agents?

  1. Reactive Agents: These are the simplest form of AI agents. They react to their environment or input without maintaining an internal state or memory. An example is a chess AI that decides moves based on the current state of the game board.

  2. Model-Based Agents: These agents have an internal model or representation of the world, which they use to make more informed decisions. They can consider the past and future states of their environment.

  3. Goal-Based Agents: These agents are designed to achieve specific goals. They evaluate their actions based on how well they contribute to achieving these goals. A self-driving car trying to reach a destination while avoiding obstacles is an example.

  4. Utility-Based Agents: These agents not only aim to achieve goals but also to maximize a given utility function. This function quantifies their performance and helps in making trade-offs between different possible actions.

  5. Learning Agents: These agents improve their performance over time based on experience. They use various learning techniques, like machine learning, to adapt to new situations and tasks.

AI agents can be found in a wide range of applications, from simple automation tasks and virtual assistants to complex decision-making systems in finance, healthcare, and beyond. Their design and implementation depend heavily on the specific requirements of the task they are built to perform.

Want some advanced AI agent info? Check out this Hugging Face paper.

Artificial GENERAL intelligence looming on the horizon?

In previous episodes we’ve talked about the techno optimists vs. pessimists, and those for and against more AI regulation.

There are moves a-foot in the regulatory sphere:

As a counter-balance, venture capitalist, Marc Andreessen, shows his techno-optimist, anti-regulatory cards. He sees AI as a force for good for the human race. His take on Responsible AI? Are we witnessing Boomer vs Doomer?

Chinese artificial intelligence developer Zhipu has raised more than 2.5 billion yuan, or $342 million, from investors since the start of the year.

How can you lead on AI in the workplace?

Roots-up initiative is great, but top-down exec engagement is key – here’s how.

Here’s some useful — and free — AI guidance for executive-types from the folks at Gartner.

Book a slot for Dec 07 — Determine not just how to use generative AI in your organization, but how to apply it to establish new competitive advantages.

It’s a top-down, exec-led task to identify the strategic opportunities for GenAI in your organization, determine what it will cost to properly utilize GenAI and assess the maturity of GenAI solutions and providers.

Uplevel your whole team with GenAI because it can level the playing field in terms of access to information and skills, making it one of the most disruptive trends of this decade.

Get involved 👍️ 👀 🙏 🧨 

We have exciting plans to develop the You & AI Newsletter and we want to involve YOU!

  • We are not just looking for feedback on You & AI News content (does it help you? Are we missing something? What do you need?).

  • We are recruiting guest authors for a new “Real world perspective” series. If you think “Nobody would be interested in what I have to say” then you’re probably wrong.

At You & AI News we want to write for as many people as possible: from their perspective; about how AI helping or hindering them; what their hopes and fears are; actionable advice to make progress and get on the AI bus instead of being under it.

If you’re interested, reply to the newsletter, share it with someone you think would be interested, or just send us an email to [email protected], or send us a message on X, or visit our LinkedIn page.