They weren’t visitors. They were agents.

You Are Now the Minority

In 2024, automated traffic surpassed human traffic on the internet for the first time in a decade. Bots now account for 51% of all web traffic. Cloudflare processes 50 billion AI crawler requests per day. GPTBot traffic alone grew 305% in one year.

The web you built for humans? Humans aren’t the primary audience anymore.

Retail sites see 59% bot traffic. Travel sites: 48%. These aren’t all scrapers or spam bots. Increasingly, they’re shopping agents, research agents, booking agents. Doing things humans used to do, on websites humans used to visit.

Cloudflare published a stat that stopped me cold. For every single visitor Anthropic refers back to a website, its crawlers have already visited 38,065 pages. OpenAI’s ratio is 1,091 to 1. Perplexity: 194 to 1. The agents read your site a thousand times for every one human they send your way.

The web hasn’t died. But it’s molting.

The Protocol War

If 2024 was the year we noticed agent traffic, 2025 was the year everyone started building the plumbing.

Anthropic released MCP (Model Context Protocol) in November 2024. People call it “USB-C for AI,” a universal adapter that lets any AI system talk to any tool or service. It now has 97 million monthly SDK downloads and over 10,000 active servers. In December 2025, Anthropic donated it to the Linux Foundation’s new Agentic AI Foundation, co-founded with Block and OpenAI. Platinum members include AWS, Google, Microsoft, Bloomberg, and Cloudflare.

Google launched A2A (Agent-to-Agent Protocol) in April 2025. It lets agents from different vendors discover each other using “Agent Cards,” basically JSON resumes. Over 150 organizations signed on, including Microsoft, Amazon, SAP, Salesforce, and PayPal. Adobe and S&P Global already use it in production.

Then the commerce-specific protocols showed up. Shopify and Google co-developed UCP (Universal Commerce Protocol), endorsed by Etsy, Wayfair, Target, and Walmart. OpenAI and Stripe built ACP (Agentic Commerce Protocol), which powers “Buy it in ChatGPT,” launched February 2026.

There’s more. Jeremy Howard proposed llms.txt, a file that tells LLMs where your best resources are (the inverse of robots.txt, which tells crawlers where NOT to go). Over 600 sites adopted it, including Anthropic, Stripe, and Cloudflare. Vercel went further, proposing embedded LLM instructions directly in HTML: <script type="text/llms.txt">. Their 401 error pages already serve agent-specific instructions.

This is the HTTP moment for agents. The protocols being written right now will shape how the agentic web works for the next decade.

When Your User Has No Eyes

We’ve spent thirty years making websites look good. Careful typography. Hero images. Hover effects. Cookie banners with the “Accept All” button slightly bigger than the “Manage Preferences” button. All designed for humans who see, click, and feel.

Your next billion users won’t see any of it.

An AI shopping agent doesn’t care about your hero image. It doesn’t notice your brand colors. It doesn’t feel the emotional pull of your “Limited Time Only” banner. It parses your structured data, checks your Schema markup, reads your JSON-LD, and makes a decision based on price, specs, availability, and reviews.

CSS is irrelevant when your user has no eyes.

Bain found that 80% of consumers already rely on zero-click results for at least 40% of their searches, reducing organic traffic by 15-25%. Google referrals to news sites dropped 9-15% in 2025. That funnel where you attract visitors with content, dazzle them with design, and convert them with psychology? Agents skip the entire thing. They go straight to the data layer.

HubSpot put it bluntly: “The fastest-growing decision-maker in your funnel cannot see your ad, feel your brand, or be persuaded by your story.”

The advertising model of the internet is about to face its first existential threat since ad blockers. Except ad blockers were opt-in. Agent browsing is default. When Perplexity’s Comet browser started bypassing Amazon’s advertising, Amazon sued. A federal judge blocked Comet from Amazon on March 10, 2026. Perplexity argued the real motivation was protecting ad revenue, not cybersecurity.

That lawsuit is a preview. The entire attention economy was built on the assumption that humans look at screens. Agents don’t look at anything.

The Money Is Already Moving

This isn’t theoretical. The money has already started flowing through agent channels.

During Cyber Week 2025, one in five orders globally were associated with AI tools or agents. That’s 20% of all orders, roughly $67 billion. On Cyber Monday alone, AI traffic to US retail sites increased 670%. AI-influenced shoppers converted 38% more frequently than traditional visitors.

McKinsey estimates agentic commerce could redirect $3-5 trillion in global retail spend by 2030, with nearly $1 trillion from the US alone. Payment executives told CNBC this could be “more transformative than the rise of e-commerce platforms such as Amazon.”

The platforms are racing to own the checkout. Shopify launched Agentic Storefronts, letting merchants appear on ChatGPT, Perplexity, Microsoft Copilot, and Google AI Mode without needing a traditional website at all. Amazon built “Buy for Me,” an AI agent that purchases from third-party brand sites so customers never leave Amazon. OpenAI launched “Buy it in ChatGPT” in February with Stripe’s Agentic Commerce Protocol behind it.

Visa launched its Trusted Agent Protocol in October 2025, an open framework to distinguish legitimate AI agents from malicious bots. Mastercard is building its own trust framework. Both are running real transactions. Not pilot stage. Deployment.

47% of US shoppers already use AI tools for at least one part of their shopping journey. That number is going one direction.

What to Do About It

The agentic web is coming whether your site is ready or not. The transition will be messy, dual-interface, and gradual. Here’s what the practical path looks like.

Structured data first. Schema markup, JSON-LD, clean OpenGraph tags. This is the content layer agents actually read. If your product pages don’t have machine-readable pricing, availability, and specs, you’re invisible to agent shoppers.

Add llms.txt. It takes ten minutes. Create a /llms.txt file that tells LLMs where your most useful resources live. Over 600 sites have done this already. It’s the new robots.txt, but instead of “go away” it says “here’s the good stuff.”

Build an MCP server. If you have an API, wrap it in MCP. Anthropic, OpenAI, Google, and Microsoft clients all support the protocol. This is how agents will interact with your service natively, without scraping your UI.

Rethink your metrics. Traffic is no longer a proxy for interest. An agent visiting your site 38,000 times doesn’t mean you have 38,000 interested customers. You need to distinguish agent traffic from human traffic and measure what agents actually do: transactions, API calls, data retrieved.

Plan for agent authentication. Visa and Mastercard are already building trust frameworks. If your business involves transactions, you’ll need a way to verify that the agent placing an order is authorized to act on behalf of a real customer.

The visual web isn’t going away tomorrow. Humans still browse. But the share of your traffic that sees your CSS is shrinking every quarter, and the share that reads your structured data is growing. Design for both.

Your Homework

Go to your website’s analytics right now. Look at your traffic. Filter for known bot user agents. The number will be higher than you expect.

Then add a llms.txt file to your site root. Ten minutes. Tell the agents where the good stuff is.

The web is being rebuilt. You can watch, or you can leave the light on for your new visitors.

They won’t see it. But they’ll know it’s there.

Claude Code might be the most significant event of 2025 in the field of AI. For me, it changed the course of the year and led me to unthinkable places. I followed every twist and turn and stayed on top of the wave. What an amazing tool that delivers on its promise. It amplifies and scales your ability, keeps you moving forward, and leads to some serious compounding.

An overlooked impact of AI is its compounding nature. From the first prompt, where everything is wrong. But, slowly, you learn, you figure out how to harness. At some point, you forget about the first hallucination, and you are in a new spot. When I talk to others, I can immediately tell where they are on the curve by how they talk about these tools. A scary thing, though, is how far behind you can get, and that you might not be able to catch up to those who have started. If you do anything this month, go down the bumpy road and get on the other side.

But today, I am thinking of myself as more of a “Todayist,” as that feels more accurate.

AI is moving faster than I thought. Predicting next week feels like sci-fi; predicting tomorrow feels ambitious. Frankly, anticipating anything beyond today is like guessing the plot of Stranger Things season 5.

Forget the grand future; I’m just trying to digest today’s AI news before tomorrow’s headlines. At this rate, “Hourlyist” is probably next. Just focusing on the now… until the next now arrives.

Empathic AI prompting is all about understanding and empathizing with your AI’s responses. Instead of getting frustrated when the AI gives a suboptimal answer, imagine why it might have produced that output. If you’ve ever done pair programming or mentored a junior developer, the idea is similar: when they make a mistake or an odd suggestion, you don’t just yell “wrong!” and walk away. You try to see what assumptions or misunderstandings led to that result, and then you guide them. With AI, you should do the same – treat it like a well-intentioned colleague who can email HR.

Why care? This mindset turns AI into a true coding partner and helps you understand more about programming. It might sound counterintuitive at first, but I am betting that this is what the future of programming looks like. It will be a required skill.

An article over at HuggingFace triggered me into this new way of thinking… Luigi says, “iterative prompting is a methodology for transforming AI into a coding partner… The key is to think of AI as a collaborator – guide it, refine its responses, and iterate until the desired results are achieved.” Another programmer, Steve Yegge from Sourcegraph even gave it a new name: CHOP (Chat-Oriented Programming) to describe this new style of coding “via iterative prompt refinement”.

Here are some of my tips to build this skill:

• Be curious, not critical: If the AI’s output is off, ask yourself why it might have thought that was okay.
• Look for clues in the response: Often the wrong answer reveals how the AI interpreted your prompt. Maybe your question was ambiguous or missing detail.
• Refine and iterate: Using those insights, clarify your prompt or provide more context, then try again.
• Collaborate step-by-step: Just like you’d work through a problem with a human, iteratively build the solution with the AI, rather than expecting perfect results first try.

This approach not only leads to better results, but also turns prompting into a learning experience. I often find that by analyzing an AI’s mistake, I understand the problem better myself! It’s a two-way street of learning.

Recently, I was working on a content-based recommendation feature for my blog. Whenever I publish a new blog post, I want to automatically recommend a few relevant forum posts from our developer community forums. Rather than writing the entire recommendation engine from scratch (with vectors, TF-IDF, and maybe cosine similarity), I decided to build this with Windsurf and its AI model’s help. Both the forum posts and blog posts are textua,l and a large language model might do a decent job of understanding which ones are related.

My first prompt to Claude Sonnet 3.5: My new blog post is about building content-based recommenders in Python, and I have a list of forum post titles. The task: pick the top 3 forum threads that would be most relevant to someone reading my blog post.

The Initial Prompt (and a Surprise in the Output)

Here’s the initial code and prompt I gave to the AI assistant (in a Python-like pseudo-code format for illustration):

blog_post = """\
Building a Content-Based Recommendation System in Python.
In this post, we explore how to recommend relevant content using textual similarity.
We use techniques like TF-IDF vectorization and cosine similarity to find related posts.
"""

forum_posts = [
    "How to implement content-based recommender with TF-IDF",
    "Collaborative filtering vs content-based: pros and cons",
    "Python 3.9 release - new features and improvements",
    "Best laptop for programming in 2025",
    "Calculating cosine similarity for text matching in Python"
]

# Initial prompt to AI
prompt = f"""
Given the new blog post:\n\"\"\"\n{blog_post}\n\"\"\"\n
and the following forum post titles:\n- {forum_posts[0]}\n- {forum_posts[1]}\n- {forum_posts[2]}\n- {forum_posts[3]}\n- {forum_posts[4]}\n
Please list the top 3 forum posts that are most relevant to the blog post's content.
"""
response = call_ai_model(prompt)  # Hypothetical function to get AI result
print(response)

When I ran this, the AI did return three forum posts, but the result made me raise an eyebrow:

1. **Python 3.9 release - new features and improvements** – (Talk about Python updates)
2. **Collaborative filtering vs content-based: pros and cons** – (Discussion comparing recommendation approaches)
3. **How to implement content-based recommender with TF-IDF** – (Guide on content-based recommendations)

At first glance, the AI’s answer isn’t horrible – it found two very relevant posts (the ones about recommender systems). But it also picked “Python 3.9 release – new features” as the top recommendation, which is clearly off-topic. A blog reader interested in content-based recommendations probably doesn’t care about the general Python 3.9 release notes. So what happened?

Empathic Analysis of the AI’s Output

Instead of cursing my AI assistant for being “dumb,” I put on my Ted Lasso hat or maybe mustache. I asked myself: Why might the AI have thought the Python 3.9 forum post was relevant?

After a moment, it clicked: the AI probably latched onto the keyword “Python” in both the blog and that forum title. My prompt didn’t stress what aspect of the content to match on, so the model might have defaulted to some simplistic notion of relevance, like overlapping keywords. The blog post text mentions “Python” (since the tutorial is in Python), and the forum post about Python 3.9 obviously mentions “Python” too – bingo, a naive content match! Meanwhile, the forum post about “Calculating cosine similarity for text matching” didn’t explicitly scream “recommender system” to the model, so it got ignored in favor of the more superficially similar Python thread.

In other words, the AI wasn’t trying to be obtuse; it was following my instructions as best it could. The mistake was actually understandable given the prompt. I realized I needed to clarify what “relevant” means in this context. This is the essence of empathic prompting: I learned something about how the AI interpreted my request, and now I can adjust my approach to guide it better.

Refining the Prompt for a Better Result

I wrote a new prompt. I wanted to tell the AI to focus on topical relevance (i.e., content about recommendation systems), and to ignore posts that are only tangentially related (like generic Python news). I also decided to explicitly mention the technique cosine similarity to nudge it toward that forum post it ignored. Essentially, I gave the AI a bit more context and a gentle steer in the right direction.

Here’s the refined prompt and the code:

# Refined prompt with more guidance
refined_prompt = f"""
We have a blog post about building a content-based recommendation system (using TF-IDF and cosine similarity in Python).
The blog post is about recommending relevant content based on text similarity.

Consider the following forum post titles:
- {forum_posts[0]}
- {forum_posts[1]}
- {forum_posts[2]}
- {forum_posts[3]}
- {forum_posts[4]}

Now, **pick the top 3 forum posts that are most relevant to the blog post's topic (content-based recommendations)**.
Focus on posts about recommendation systems or text similarity (TF-IDF, cosine similarity), *not* general Python news or unrelated topics.
"""
response = call_ai_model(refined_prompt)
print(response)

And the AI’s new response:

1. **How to implement content-based recommender with TF-IDF** – A forum thread directly about content-based recommendation using TF-IDF, which is exactly the technique discussed in the blog.
2. **Collaborative filtering vs content-based: pros and cons** – A discussion comparing content-based methods with collaborative filtering, which complements the blog post's topic.
3. **Calculating cosine similarity for text matching in Python** – A thread about cosine similarity (a core part of content-based recommendations), very relevant to the blog post.

The irrelevant Python 3.9 post is gone, replaced by the cosine similarity thread that we actually want. Notice how I didn’t tell the AI explicitly which ones to choose; I simply guided it by clarifying what “relevant” really means here. The AI did the rest on its own once it understood my intentions better.

A little empathy went a long way: by imagining what the AI was thinking before and adjusting my wording, I got a much better result. In a way, it felt like I debugged my prompt by examining the AI’s “mistake” and then fixing the cause of that mistake.

Let Me Try Explaining This Again

• Think of the AI as a teammate. It’s not a perfect genius that will always get everything right, nor is it a dumb tool to be dismissed when it errs. Like any human collaborator, it has its own way of “thinking” (albeit a very alien, statistical one!). Engage with it, guide it, and be patient when it needs a second try. As I reminded myself, treat AI-generated code or answers as you would a human junior developer’s output. You’d review it, figure out reasoning, and help improve it – do the same with the AI output.
• Use mistakes as insights. When the AI goes off track, don’t just reject the output – inspect it. Is there a pattern to the misunderstanding? Maybe your question was ambiguous or assumed knowledge the AI didn’t have. In our case, “relevant” was too vague without context. Often, an AI’s “mistake” is really a mirror held up to our prompt: it reflects how our request could be interpreted. This is incredibly useful for honing our own communication and problem decomposition skills.
• Iterate and refine. The first answer is rarely the final answer. Professional developers know that software design is an iterative process – turns out, prompt design is iterative too. You might start with a rough prompt, see what the AI does, then refine the prompt multiple times. This iterative loop is at the heart of chat-oriented programming (CHOP) and effective prompt engineering. It’s not cheating or a failure to redo the prompt; it’s part of the workflow. Each iteration is a chance to incorporate what you learned from the last output.
• Keep a casual, curious mindset. One thing I enjoy about working with an AI is that I can think out loud. If the AI says something weird, I can literally ask it “why did you do that?” or I can just introspect and adjust. Treating it like a game or puzzle is a bit of fun – “hmm, the AI misunderstood me, how can I say this differently?” Maintaining that curiosity (and even humor) makes the process less frustrating. Remember, the AI has a lot more patience than we do and doesn’t mind if you rephrase the question five times before you get it right

Empathic prompting is a skill that takes time to develop. You have to consistently reset your expectations as models get better and IDEs improve, but it pays off.

Articles and Videos that you need to read/watch, or AI summarize:

• The Right Way to Use AI for Coding (And How It Transformed My Workflow)
• Chat-oriented programming (CHOP) in action
• Mastering Iterative Prompting for Optimized AI Code Generation

TL;DR

GTC Keynote from Jensen Huang, CEO of NVIDIA

Jensen Huang, CEO of NVIDIA, at GTC March 2024 discusses the evolution of computing and the emergence of generative AI, a new kind of software that can be used to create other software. He also unveils a new line of products from NVIDIA, including NIM (NVIDIA Inference Microservice), and Omniverse, a digital twin platform that can be used to create and simulate robotic systems.

• Generative AI is a new kind of software that can be used to create other software. This is a major breakthrough that has the potential to revolutionize many industries.
• NVIDIA is a leader in the development of generative AI, and the company is committed to making this technology accessible to everyone. Nims is a new product from NVIDIA that makes it easy for companies to integrate generative AI into their workflows.
• Omniverse is another new product from NVIDIA that is designed to help companies create and simulate robotic systems. Omniverse is a digital twin platform that can be used to create virtual models of real-world systems. These models can then be used to train robots and to test new products and processes.
• NVIDIA believes that generative AI and Omniverse have the potential to transform many industries, including healthcare, manufacturing, and transportation. The company is committed to working with its partners to bring these technologies to market.

The Power of the Ecosystem

The one awesome part of GTC is that NVIDIA is bringing together partners and building a strong ecosystem. There have been so many blogs, press releases, and videos to capture all of the news, but one is the idea that it is going to take many developers, partners, and tech stacks to come together in order to deliver on the promise of AI. I was proud to see that MathWorks is in the ecosystem.

One thing that has captured my interest is the Omniverse Cloud APIs, which I believe are set to revolutionize autonomous system development through advanced sensor simulation. These APIs offer developers crucial tools for creating and testing realistic scenarios, thus enhancing the safety and effectiveness of autonomous technologies. NVIDIA’s initiative not only accelerates the development process but also aims to democratize access to cutting-edge technologies, fostering innovation in the autonomous sector.

By bringing together an expansive ecosystem of simulators, verification and validation (V&V) tools, content and sensor developers, the Omniverse Cloud APIs enable a universal environment for AI system development.

Developers and software vendors such as CARLA, MathWorks, MITRE, Foretellix, and Voxel51 underscore the broad appeal of these APIs for autonomous vehicles.

Here’s a video showing how Foretellix leverages NVIDIA Omniverse Cloud APIs to generate high-fidelity sensor simulations for autonomous vehicle development.

2024 is shaping up to be exciting. AI tech, platforms, partners, and developer interest are hitting at once.

My previous attempts worked out well, but I recently found out that AutoHotkey has changed their configuration files and the project now defaults to AutoHotkey v2 with AutoHotkey v1/v1.1 being deprecated. I have updated my tutorial to AutoHotkey v2.

Big picture. We are going to install the Norwii Presenter software, configure the buttons on the presenter remote, and create a custom AutoHotkey v2 script to intercept the button presses and send them to PowerPoint. If you don’t do this, the presenter remote will only work when PowerPoint is in focus. That works well in some situations, but it doesn’t work as well for online meetings where you are moving around your desktop to monitor the chat, look a response, or find a document.

Buy a Norwii Presenter Remote on Amazon. I like the Norwii N28 model with its not easy to miss big button for advancing slides, but the other models are lovely too. The N28 is wireless and most buttons are customizable. I also like this model since it uses a AAA battery instead of a rechargeable battery. The rechargeable model sounds like a good idea, but I have found out that it is always dead when you want to use it. With the replaceable battery model, you just replace the battery if it is dead. From my experience and usage, the AAA battery lasts almost a year.

Install the Norwii Presenter Utility for Windows. This software has the driver for the wireless dongle and also a tool for remapping the buttons of your remote to new buttons. You can remap all of the buttons (except the laser pointer button). The software is not necessarily easy to use, but it gets the job done. Other presentation remotes lack the ability to be reprogrammed. This is why I love the Norwii models.

Customize the Norwii Presenter buttons. Click on the Customize tab in the utility and make sure N28 (M) is selected as the model. Under each key configure the short press function for each mapping that you have in AutoHotkey (more on that in a moment). The big yellow button is known as the “Page down” key. I use Shift + F12 to advance my slides in PowerPoint. you can do the other buttons if you like. I have experimented with programing all of the buttons to advance the slides. That way I don’t have to fumble around trying to pick the right button in the heat of the moment. But, I have expanded this to backing up a slide (page up key) and exiting the presentation with the enter key. I practiced a few times to get this right. Let me know in the comments what works for you.

Also, make sure that the Norwii Presenter Utility starts up with Windows. This will make sure that the utility is running in the background after you reboot to install Windows updates. Click the Setup tab and select, “Start Norwii Presenter after Windows startup.”

Install and configure AutoHotkey. We are going to use an awesome utility for Windows PCs called AutoHotkey. This utility listen for your custom key presses redirects the controls to PowerPoint. Let me warn you now ff this is your first time using AutoHotkey. You are going to get addicted to automation. YOU HAVE BEEN WARNED.

Install AutoHotkey v2 on your Windows PC. AHK will run in the background waiting for keyboard commands that match the rules that you have created. For help, check out the documentation and the Hello World video on YouTube. I know, it will be daunting at first, but a community of people for the last decade have been creating videos, examples, and tutorials. You just might be inspired to automate other computing tasks. My goal is to map Shift+F12 and Shift+F11 to Page Up and Page Down. Page Up and Page Down are keyboard shortcuts for PowerPoint.

1. Right-click on your desktop
2. Find “New” in the menu
3. Click “AutoHotkey Script” inside the “New” menu
4. Name it PowerPoint.ahk
5. Right-click on PowerPoint.ahk and click “Edit Script”
6. Enter the this AutoHotkey 2 code and save the file:

#Requires AutoHotkey v2.0

;; Key Bindings for AutoHotkey v2 and PowerPoint
;; nothans.com/powerpoint

; Bind to Shift+F12 - Advance slide or start presenting from the last slide
+F12::
{    
    if WinExist("PowerPoint Slide Show")
    {
        WinActivate
        Send "{PgDn}"
        return
    }
    else
    {
        if WinExist("ahk_exe POWERPNT.EXE") {
             WinActivate
	     Send "+{F5}"	
        }        
        return
    }
}

; Bind to Shift+F11 - Go back one slide
+F11::
{
    if WinExist("PowerPoint Slide Show")
    {
        WinActivate
        Send "{PgUp}"
        return
    }

    return
}

; Bind to Shift+F8 - Exit slideshow
+F8::
{       
    if WinExist("PowerPoint Slide Show")
    {
        WinActivate
        Send "{Esc}"
	return
    }

    return
}

; Bind to Shift+F7 - Advance slide or start presenting from the first slide
+F7::
{    
    if WinExist("PowerPoint Slide Show")
    {
        WinActivate
        Send "{PgDn}"
        return
    }
    else
    {
        if WinExist("ahk_exe POWERPNT.EXE") {
             WinActivate
	     Send "{F5}"	
        }        
        return
    }
}

For more AutoHotkey button mappings, check out my GitHub repository.

Use your new power to deliver the ultimate PowerPoint presentation. Now that you are all powerful, open PowerPoint and click the big yellow button on the Norwii N28 remote. This will put your presentation into presentation mode. Click the big yellow button again. Your slides or animations advance. Click it again. Your slides advance. It is a beautiful thing.

Elgato Stream Deck Control

You can use the same AutoHotkey v2 configuration to control PowerPoint from the Elgato Stream Deck or Elgato Stream Deck Pedal as well.

1. Open up the Stream Deck Software and search for “Hotkey”

2. Drag the Hotkey option to an open button

3. Click into the Hotkey field and wait for it say, “Observing keystrokes…”, and press Shift+F12

4. Rename the title to, “Advance”, and set the icon to the forward icon

5. Create Stream Deck hotkeys for the rest of the key bindings in your AutoHotkey v2 configuration and enjoy the power!

Let me know in the comments if you try this out. Also, let me know if you have some tricks for PowerPoint or other Windows apps. I am always looking to optimize.

Microsoft Build 2023 TL;DR

1. Bing Integration with ChatGPT: Bing has been integrated into ChatGPT as the default search engine, allowing for real-time queries with up-to-date results, fresh content, and citations. Users can directly access webpages sourced by Bing from these results.
2. Plugins in Bing and ChatGPT: The presentation announced a new interoperability feature between ChatGPT and Bing for plugins. The same plugins can now be used in both platforms, and an example was given of using Zillow and Instacart plugins to search for houses and create a shopping list.
3. Edge Browser and Bing Chat: The Edge browser now supports Bing Chat, allowing it to pull information, like ingredients from a recipe on a webpage, and generate a shopping list using the Instacart plugin.
4. Windows Copilot: A new feature called Windows Copilot was introduced. This tool gives users suggestions to increase productivity and enhance user experience on a PC. The tool allows users to adjust system settings, switch to focus or dark mode, and manage window arrangement with one click. It can also invoke plugins, such as playing a playlist from Spotify.
5. Microsoft Fabric: Microsoft announced a unified data analytics platform called Microsoft Fabric. It combines compute and storage, integrates the full analytics stack product experiences, unifies governance with analytics, and brings a unified business model across all analytics workloads. The platform includes OneLake, a SaaS data lake that stores data in an open format for diverse use cases like training machine learning models, visualizing data, and running SQL queries. The unified experience provides data professionals with all the tools they need.

I admit it freely. I love new tech, but I don’t jump all of the hype trains. AI has swept me along and Mojo definitely struck me as something to pay attention to. I am in constant awe caused by the rapid advancements in this field. It seems like every time I blink, a new breakthrough is making waves. Today, it’s a new programming language called Mojo, developed by Modular.

Mojo is an intriguing blend of Python’s syntax and ecosystem, coupled with the robust features of systems programming and metaprogramming. The language is in its infancy and is evolving rapidly. It is designed to be a superset of Python over time. This means developers like me can customize and squeeze the most out of our hardware.

What’s particularly cool about Mojo is how it combines the accessibility of Python with the power of systems programming. It uses MLIR to scale to new and exotic hardware types and domains, a feat that other languages haven’t yet accomplished. Autotuning, caching, and distributed compilation are all built right into its core. Mojo is also shaking things up with its unique approach to memory ownership and management, and its ability to scale down to smaller envelopes. It’s Python-first, which means I don’t have to learn an entirely new syntax.

As a general-purpose programming language, Mojo isn’t just for AI. It can be used for a multitude of tasks, including high-performance computing (HPC), data transformations, and writing pre/post-processing operations. It’s set to support more architectures over time, and it even includes a debugger and a full tool suite. As a superset of Python, Mojo is designed to let developers write portable code that’s faster than C, and it allows for seamless interoperability with the Python ecosystem.

Mojo is still in the early stages of development, so only simple programs can be transitioned over as-is without any code changes. But as the language matures, migration tools will be developed. The developers of Mojo are keen on enabling developers to port code from languages other than Python to Mojo, and they’re even eyeing migration from C/C++ due to Mojo’s similarity to the C/C++ type systems.

The developers plan to progressively open-source Mojo over time, but for now, it’s changing rapidly and being incubated within Modular. As a new language, there aren’t any widely known AI-related performance benchmarks for Mojo yet, but Modular’s in-house kernel for the Modular Inference Engine is written entirely in Mojo.

Mandelbrot in Mojo

Mojo is not just exceptional for crafting high-performance code, but it also offers the ability to tap into the vast ecosystem of Python’s libraries and tools. Thanks to seamless Python interoperability, Mojo can utilize Python’s strengths, particularly in GUIs, without compromising performance in crucial code segments.

Let’s see how this might work working with the Mandelbrot Set in Mojo.

You can first take the complex code to generate the Mandelbrot set in Mojo and vectorize it using the Functional library. Once the code is optimized, you can take the code and parallelize it using the Functional library. This will increase the performance by parallelizing on the rows.

You sign up to try the Mojo Playground on the Modular website.

I will share my experience generating art with AI. I am not an artist, but I have ideas and access to a web browser that has access to power Generative AI tech. There are many platforms to experience, but I recommend that you pick one, try out prompts, iterate, and share. I picked Stable Diffusion.

Stable Diffusion is an advanced generative model for creating images, inspired by the diffusion process in physics. When given a text prompt, the model generates an image based on the content and themes described in the prompt.

M. C. Loves Robots

I love the artstyle of M. C. Escher. With Generative AI, I can now get an idea of what Escher would create if he was obsessed with robots.

Let’s unpack the prompt used to generate this image

The prompt “M. C. Escher lithograph, robots, gears, lights, wires” consists of several components that give the model an idea of what elements to include in the generated image.

1. “M. C. Escher lithograph” – This part of the prompt instructs the model to generate an image inspired by the artistic style of M.C. Escher, a Dutch graphic artist known for his mathematically inspired works that feature impossible objects, optical illusions, and intricate patterns.
2. “robots” – This keyword tells the model to include robots as one of the central subjects in the image. The robots may be designed in a way that reflects Escher’s style or is inspired by his works.
3. “gears” – The model should incorporate gears into the image, possibly as part of the robots’ mechanisms or as a separate visual element.
4. “lights” – This term suggests that the image should have a focus on lighting, either by using the lights as a part of the composition or emphasizing the way they interact with the other elements in the scene.
5. “wires” – The presence of wires indicates that the generated image should have a somewhat intricate or complex appearance, with wires possibly connecting various elements or being part of the robots.

When the Stable Diffusion model receives this prompt, it will generate an image that combines these elements. The resulting artwork might be an intricate scene of robots, gears, and wires, all rendered in the distinctive style of M.C. Escher, with a focus on the interplay of light and shadow. The specific details of the generated image will depend on the model’s interpretation of the prompt and its internal knowledge of the various elements mentioned.

Each word in a Generative AI prompt matters. You can get wild variations by changing out the words and experimenting.

Stable Diffusion Prompt Engineering Tips

Stable Diffusion is a groundbreaking generative AI art model that enables the creation of stunning and unique artwork. By providing the model with a carefully designed prompt, you can guide its creative process to generate art that aligns with your vision.

Be Specific and Clear

When writing prompts for Stable Diffusion, it’s crucial to be specific and clear in your description. This ensures that the model can accurately interpret your intentions and generate artwork that matches your vision. For example, instead of writing “a landscape,” provide more detail such as “a serene sunset landscape with a calm lake and a silhouette of a tree.”

Incorporate Emotions and Atmosphere

To create truly captivating art, consider incorporating emotions and atmosphere into your prompts. This can help the model generate art that evokes a certain mood or feeling. For instance, instead of writing “a cityscape,” try “a bustling cityscape with a lively atmosphere, filled with people enjoying their day.”

Experiment with Different Styles

Stable Diffusion is capable of producing artwork in various styles, from abstract to photorealistic. When crafting your prompt, consider specifying a particular artistic style or movement to inspire the model. For example, you could write “a vibrant, impressionistic painting of a Parisian street café.”

Combine Unusual Elements

One of the most exciting aspects of generative AI art is its ability to combine seemingly unrelated elements to create unique and thought-provoking pieces. To harness this potential, try writing prompts that incorporate contrasting or unexpected components. For example, “a surreal scene of a whale floating above a city skyline with hot air balloons in the background.”

Iterative Refinement

Don’t be afraid to iterate and refine your prompts. If the generated artwork isn’t quite what you envisioned, try adjusting the wording or adding more details to your prompt. It’s essential to experiment and adapt your prompts to hone your skills in guiding the model’s creativity.

Stable Diffusion offers an exciting avenue for artistic expression, and writing effective prompts is key to unlocking its full potential. By being specific, incorporating emotions, experimenting with styles, combining unusual elements, and refining your prompts, you can guide the model to generate visually stunning and emotionally evocative artwork. So go ahead, unleash your creativity, and let Stable Diffusion bring your artistic visions to life!

Use ChatGPT if you get stuck

A neat thing that you can do is use ChatGPT if you get stuck. Use AI to write AI prompts. Ask ChatGPT to give you ideas for prompts for an artist that you like. I always loved President Obama’s presidential painting by Kehinde Wiley. A robot is likely to be our future president and Kehinde will be tapped once again to paint the portrait. I used ChatGPT to help me generate ideas for this painting.

Enjoy the journey

We are going through some weird times. Enjoy the journey. Try out prompts and share the results on our Discord server. I would love to see what you come up with.

AI Name: Blogger-AI

Blogger-AI is: An AI that writes an 8000-word blog post about Auto-GPT

Goals:

• Goal 1: write intro paragraph: tell me what it is
• Goal 2: write a paragraph: comparing autogpt vs chatgpt
• Goal 3: write more paragraphs at your direction
• Goal 4: end with a conclusion and CTA to join my discord server

After five minutes and Auto-GPT trying to install Auto-GPT so it could better explain Auto-GPT, this is the blog post that it wrote. Enjoy.

What is AutoGPT?

AutoGPT is an AI tool that uses GPT-3.5 technology to generate human-like text. It is designed to be used in a variety of contexts, including content creation, customer service, and more. AutoGPT is different from ChatGPT in that it is more focused on generating text for specific purposes, rather than engaging in open-ended conversations. With AutoGPT, users can generate high-quality text quickly and easily, making it a valuable tool for businesses and individuals alike.

AutoGPT is an open source project that aims to provide an autonomous version of GPT. The key features of AutoGPT include internet access for searches and information gathering, long-term and short-term memory management, GPT-4 instances for text generation, access to popular websites and platforms, file storage and summarization with GPT-3.5, and the requirement of Python 3.8 or later and an OpenAI API key. Optional features include an ElevenLabs Key for speech mode and the ability to use GPT on an Azure instance. The installation process involves cloning the repository, navigating to the project directory, installing required dependencies, and filling in API keys. Usage involves running the main.py Python script in the terminal and authorizing each action with ‘NEXT COMMAND.’ Logs can be found in the ./logs folder and debug logs can be outputted with a flag. Other features include speech mode, Google API key configuration, Redis setup, Pinecone API key setup, and the ability to switch between local cache, Pinecone, and Redis for memory storage. Limitations include the experiment being an unpolished application and potentially not performing well in complex business scenarios.

How Does AutoGPT Work?

AutoGPT works by combining GPT with a companion robot that uses GPT and several APIs to achieve a set goal. The companion robot is responsible for making decisions and taking actions based on the output of GPT. The robot uses a combination of APIs, including Google Search, to gather information and make decisions. AutoGPT also includes long-term and short-term memory management, which allows the system to remember previous actions and decisions. The system uses GPT-4 instances for text generation, which allows it to generate high-quality text in a variety of styles and formats. AutoGPT also includes access to popular websites and platforms, which allows it to perform a wide range of tasks. The system uses file storage and summarization with GPT-3.5, which allows it to store and summarize large amounts of information. The installation process for AutoGPT involves cloning the repository, navigating to the project directory, installing required dependencies, and filling in API keys. Usage involves running the main.py Python script in the terminal and authorizing each action with ‘NEXT COMMAND.’ Logs can be found in the ./logs folder and debug logs can be outputted with a flag. Other features of AutoGPT include speech mode, Google API key configuration, Redis setup, Pinecone API key setup, and the ability to switch between local cache, Pinecone, and Redis for memory storage. Limitations of AutoGPT include the experiment being an unpolished application and potentially not performing well in complex business scenarios.

AutoGPT is a powerful language model that is designed for automated text generation. It is built on top of the GPT architecture and is capable of generating high-quality text with minimal input from the user.

AutoGPT Features

One of the key features of AutoGPT is its ability to generate text in a more automated fashion. This means that it can generate text without requiring a human agent to prompt its every action. This makes it ideal for content creation, summarization, and translation tasks.

Another key feature of AutoGPT is its ability to generate coherent and high-quality text. It is trained on a large corpus of text data, which allows it to generate text that is both grammatically correct and semantically meaningful. This makes it suitable for a wide range of applications, including content creation, chatbots, and customer service.

Overall, AutoGPT is a powerful tool that can be used in a variety of contexts. Its ability to generate high-quality text with minimal input from the user makes it ideal for content creation and other automated text generation tasks. Its ability to understand the context of the conversation and generate appropriate responses makes it suitable for chatbot and customer service applications.

AutoGPT Use Cases

AutoGPT can be used in a variety of applications and use cases. One of the most common applications of AutoGPT is content creation. It can be used to generate articles, blog posts, and other types of content with minimal input from the user. This can save content creators a significant amount of time and effort, allowing them to focus on other aspects of their work.

Another common use case for AutoGPT is summarization. It can be used to summarize long articles or documents into shorter, more digestible pieces of content. This can be useful for people who need to quickly understand the key points of a document without reading the entire thing.

AutoGPT can also be used for translation. It can be trained on multiple languages, allowing it to generate high-quality translations with minimal input from the user. This can be useful for businesses that need to translate large amounts of content quickly and accurately.

Overall, AutoGPT is a versatile tool that can be used in a variety of applications and use cases. Its ability to generate high-quality text with minimal input from the user makes it a valuable tool for content creators, summarization, and translation tasks.

Conclusion

In conclusion, AutoGPT is a powerful tool for generating high-quality text with minimal input from the user. It can be used in a variety of applications and use cases, including content creation, summarization, and translation. While it has some limitations, such as its inability to engage in a conversation with the user, it is still a valuable tool for anyone who needs to generate high-quality text quickly and efficiently.

If you are interested in learning more about AutoGPT and other AI tools, please consider joining my Discord server. We have a community of AI enthusiasts who are always happy to share their knowledge and expertise. Thank you for reading!

Bonus

At the end of the blog post writing it came up with the idea to tweet the blog post as a series of tweets. I didn’t have that feature enabled, but it tried to do it in a variety of ways. It created a new file with Twitter credentials and wrote a Python script to try and automated the tweets. It tried five different ways before I told it to stop.