I’m rebranding my other newsletter (Product Thinking) and combining it with this newsletter (Next Horizon) into a single, kick ass newsletter called Prodity. I’ll continue to focus on AI, technology, product management, and design, but be able to focus more on bringing the best of everything in one spot.

What to do?

Since you are a paid subscriber, I…

Read more

I’m rebranding my other newsletter (Product Thinking) and combining it with this newsletter (Next Horizon) into a single, kick ass newsletter called Prodity. I’ll continue to focus on product management, design, technology, and AI, but be able to focus more on bringing the best of everything in one spot.

What to do?

Nothing.

If you are a subscriber, I’ve already combined everything for you. You don’t need to do anything. You’ll keep getting the newsletters as always. Just keep reading and sharing now that you get the combined power of both! If you are new and reading this one, got subscribe to Prodity.

What will it be?

• Free weekly update

• 2-3 deep-dives each month (paid subscribers)

We’ll be adjusting the structure slightly (we’re always experimenting like product managers do). The plan now is to do a weekly recap of the best information of the week in product management, AI, and technology.

For our paid members, we’ll be continuing our deeper dives 2-3 times per month. This includes exploring product roadmaps, case studies, implications of AI, and book reviews. So be sure to upgrade for in-depth content.

Subscribed

We also have our Prodity podcast, Product by Design. I’ll keep having engaging and insightful conversations with founders, CEOs, designers, product managers, and more. Follow the podcast so you don’t miss out!

Why?

Focus.

I’ve been creating a lot of content across several places and need to bring everything together to focus. I’ve been thinking about this for a long time as my newsletters and podcasts have grown (both in number and size), and am incredibly excited to bring everything together into something more cohesive.

I hope you continue to enjoy my thoughts and updates as I reflect on technology, product management, design, and ideas that can help you in your career and life.

• 🦙 Llama

• 🖥️ Apple

• ✖️ Elon

• 🖼️ Images

• ✒️ AI content

• 😺 AI Optimism

Latest News and Updates

Meta Releases Llama 2

Meta released Llama 2 this week, an open-source large language model (LLM). Unlike other LLMs, they aren’t keeping anything hidden. The code is available to anyone for research or commercial purposes. The hope is to drive innovation forward, along with better responsibility:

“Open-source drives innovation because it enables many more developers to build with new technology,” Zuckerberg said in a Facebook post. “It also improves safety and security because when software is open, more people can scrutinize it to identify and fix potential issues.”

If you want to start using Llama 2, you can try many sites, like Llama2.ai and Llama.perplexity.ai (which seems faster in my testing).

Apple Tests ‘Apple GPT,’ Develops Generative AI Tools to Catch OpenAI

Apple recently banned its employees from using AI tools. Unsurprisingly, it is creating its own:

Apple Inc. is quietly working on artificial intelligence tools that could challenge those of OpenAI Inc., Alphabet Inc.’s Google and others, but the company has yet to devise a clear strategy for releasing the technology to consumers.

Elon Musk unveils his AI company, xAI

Not to be outdone, Elon also launched an AI company:

On Wednesday, an official website for Musk's AI venture went live at X.AI. A Twitter account also launched with the handle @xAI and affiliate badges for individuals working at the newly formed company. Musk then confirmed his involvement by changing his own personal Twitter profile's bio to simply the @xAI handle.

Report: OpenAI holding back GPT-4 image features on fears of privacy issues

Some AI tools have image features embedded, and others are stand-alone image generators. I know I’ve been waiting for images in ChatGPT for some time. But they are holding back still:

OpenAI has been testing its multimodal version of GPT-4 with image-recognition support prior to a planned wide release. However, public access is being curtailed due to concerns about its ability to potentially recognize specific individuals, according to a New York Times report on Tuesday.

Useful Tools & Resources

Copy Leaks

I use ChatGPT, Bard, and other AI tools to help with many things. From generating ideas to refining content to brainstorming. What I don’t want, though, is to have too much AI generated content in anything I do. So being able to test my own content, and other things I see on the internet, is very helpful.

CopyLeaks does a great job at detecting AI generated content vs human-generated content. I expect eventually these kinds of tools will be incorporated in everything we do, but for now, at least there is a place to check.

Deep Dive - Anthropomorphizing

I was having a conversation recently about the anthropomorphizing of AI. It has happened since the beginning of AI and chatbots, and we’ll likely see it continue. This article on AI love: It's complicated dives into more:

Now, some are finding love with an artificial intelligence. The online forum Reddit is filled with people confessing their love for AI chatbots, such as Replika. 

But is it real love? To understand whether this "digital love" is the same or similar to the love we experience with other humans is tough. Because it's hard enough to understand what "normal" love is.

This will be a fascinating area to observe and study, but one that likely become a bigger part of life and society…

Share

Subscribe now

• 🤖 Robots

• 🖥️ Supercomputers

• ↔️ Alignment Problem

• 📝 AI Terms

• ✒️ Prompting

• 😺 AI Optimism

Latest News and Updates

DeepMind’s New Self-Improving Robot Is Quick to Adapt and Learn Fresh Skills

The progress in intelligent robots has continued forward, though not as quickly as many of us may have expected. Robots are still not great, but we’ll see that begin to change with all the other advances:

While autonomous robots have started to move out of the lab and into the real world, they remain fragile. Slight changes in the environment or lighting conditions can easily throw off the AI that controls them, and these models have to be extensively trained on specific hardware configurations before they can carry out useful tasks.

This lies in stark contrast to the latest LLMs, which have proven adept at generalizing their skills to a broad range of tasks, often in unfamiliar contexts. That’s prompted growing interest in seeing whether the underlying technology—an architecture known as a transformer—could lead to breakthroughs in robotics.

Amazon has also been making progress on its robotic automations:

Proteus is part of an army of smarter robots currently rolling into Amazon’s already heavily automated fulfillment centers. Some of these machines, such as Proteus, will work among humans. And many of them take on tasks previously done by people.

Inflection AI Develops Supercomputer Equipped With 22,000 NVIDIA H100 AI GPUs

We’re seeing the beginning of AI supercomputers. Many more will likely follow. Which raises many questions, like how will power all these and how soon will we need to create a Dyson sphere around our sun just for energy.

Inflection AI's Supercomputer Expected to Be One Of The Largest In the Industry, Right Behind AMD's Frontier

For those unfamiliar with Inflection AI, it is a business that aims at creating "personal AI for everyone." The company is widely known for its recently introduced Inflection-1 AI model, which powers the Pi chatbot. Although the AI model hasn't yet reached the level of ChatGPT or Google's LaMDA models, reports suggest that Inflection-1 performs well on "common sense" tasks, making it much more suitable for applications such as personal assistance.

OpenAI Tackles Superalignment

The alignment problem may be one of the most important challenges in AI. I discussed this in another post:

Product Thinking

The Alignment Problem: Machine Learning and Human Values

The book The Alignment Problem by Brian Christian has been on my reading list for a while now. It came up again in a Product by Design podcast interview I was doing as we discussed the implications of AI, good and bad. And with my own increased interest in AI (don’t forget to read and share the…

Read more

3 years ago · Kyle Evans

OpenAI is dedicating a lot of resources to ensuring that the AI we create, which could be smarter than us, will remain aligned:

Superintelligence will be the most impactful technology humanity has ever invented, and could help us solve many of the world’s most important problems. But the vast power of superintelligence could also be very dangerous, and could lead to the disempowerment of humanity or even human extinction.

Currently, we don't have a solution for steering or controlling a potentially superintelligent AI, and preventing it from going rogue. Our current techniques for aligning AI, such as reinforcement learning from human feedback, rely on humans’ ability to supervise AI. But humans won’t be able to reliably supervise AI systems much smarter than us.

Artificial intelligence glossary: 60+ terms to know

As this article states:

It is getting harder to keep up with AI and the growing list of jargon and scientific terms surrounding it…

This glossary aims to serve both as a resource for those just being introduced to AI and for those looking for a reference or to refresh their vocabulary.

A useful reference for the growing list of terms.

Useful Tools & Resources

This week I explored tools to help with prompt generation. This is a broad topic, so we’ll have more to come. I’ll be creating a library of these tools soon, so remember to check back for everything soon.

I’ve been experimenting with FlowGPT to see prompts and generate prompts. It has a ton of existing prompts that you can use or customize. Prompt engineering is becoming a critical part of using AI products. So tools like these are helpful, even just to see how others are successfully creating prompts.

A redditor also experimented with a few tools and ranked them, which I found interesting. You can check it out here.

Deep Dive - The Optimistic Case

We continue to explore the many opportunities and challenges that AI will present. There is incredible reason to be optimistic, but also many reasons to worry. Those aren’t mutually exclusive.

In an article in The Guardian, they explore both sides. We’ll focus on the optimistic case right now:

Recent advances such as Open AI’s GPT-4 chatbot have awakened the world to how sophisticated artificial intelligence has become and how rapidly the field is advancing. Could this powerful new technology help save the world? We asked five leading AI researchers to lay out their best-case scenarios.

First, AI will have the ability to improve everything about the human experience. We will be able to cure more diseases, extend our lives, and become better versions of ourselves:

Most movies about AI have an “us versus them” mentality, but that’s really not the case. This is not an alien invasion of intelligent machines; it’s the result of our own efforts to make our infrastructure and our way of life more intelligent. It’s part of human endeavour. We merge with our machines. Ultimately, they will extend who we are. Our mobile phone, for example, makes us more intelligent and able to communicate with each other. It’s really part of us already. It might not be literally connected to you, but nobody leaves home without one. It’s like half your brain.

AI will be able to take all the information and complexities of the human body and help us understand them and their interactions. Ultimately revolutionizing how we treat and care for ourselves:

One reason AI can be useful here is that the body is very complicated. Even a single cell is extremely complicated: you have 20,000 genes, and they all interact with each other. Biotechnology has progressed to the point where we can measure all the genes’ activity in a single cell at once. While we collect huge quantities of data, the quantity of data is so large that humans are unable to read it. But because machines can, they are able to build models of how your cells work, and how they could be changing under different circumstances that cause disease. So, you can see what happens if you make an intervention; if you introduce a pollutant or a drug, what will be the effect?

We’ll also be able to accelerate the improvements of technology. The tools we’re creating could create additional tools and innovations that accelerate progress in a way that has been unimaginable.

In just a couple of decades, humanity could get to the kind of advanced future that feels like it’s hundreds or thousands of years away. This is not at all guaranteed, but I think it’s within reach if we get this right.

Ultimately, our innovations will allow us to continue to innovate and flourish, which I find very hopeful:

We can cure all diseases, stabilise our climate, eliminate poverty, etc. We can flourish not just for the next election cycle, but for billions of years. We have been on this planet for more than 100,000 years, and most of the time we have been like a leaf blowing around in the wind, without much control of our destiny, just trying to not starve or get eaten. Science and technology and human intelligence have made us the captains of our own ship.

Share

Subscribe now

Here’s this week’s news, product applications, and broader philosophical implications. So here we go:

• 🖼️ Midjourney

• 🧮 AI Survey

• 🚜 Farming

• 🤖 LLM Agents

• ⛺ Camping

• ✒️ Marc Andreessen

Latest News and Updates

“Stunning”—Midjourney update wows AI artists with camera-like feature

Midjourney continues to push the boundaries of AI image generation:

Midjourney unveiled version 5.2 of its AI-powered image synthesis model, which includes a new "zoom out" feature that allows maintaining a central synthesized image while automatically building out a larger scene around it, simulating zooming out with a camera lens.

Hope, fear, and AI

We previously discussed a survey from March 2023 that showed how ChatGPT was still relatively unknown to most Americans. As of June 2023, we can see that more people are getting familiar with AI. But it is still fewer than you might think.

But despite widespread news coverage, use of these new tools is still fairly limited, at least when it comes to dedicated AI products. And experience with these tools skews decidedly toward younger users.

@ai.nexthorizonwho has heard of chatGPT and other tools? fewer people than you might think #ai #chatgpt #greenscreen

Tiktok failed to load.

Enable 3rd party cookies or use another browser

AI comes to the farm: New technology tackles acres of weeds quickly by using robotics and lasers

It’s not just those of us in offices who will benefit (or be impacted) by AI:

Farmers interested in a fast, accurate way to rid their fields of weeds have a new option in the AI space. Carbon Robotics is now shipping its LaserWeeder to farms around the United States; the machine uses the power of lasers and robotics to rid fields of weeds.

Weeds are one of the most "tedious, time-consuming and challenging" elements of farming, Carbon Robotics told Fox Business via email. 

LLM Powered Autonomous Agents

Looking for more info about LLM agents? Here is a good resource:

Building agents with LLM (large language model) as its core controller is a cool concept. Several proof-of-concepts demos, such as AutoGPT, GPT-Engineer and BabAGI, serve as inspiring examples. The potentiality of LLM extends beyond generating well-written copies, stories, essays and programs; it can be framed as a powerful general problem solver.

Useful Tools & Resources

This week I explored tools to help plan camping trips. I’ll be creating a library of these tools soon, so remember to check back for everything soon.

AdventureGenie

We’re going camping soon, so naturally I thought it would be great to see what kinds of tools are available to help plan and manage trips. I haven’t subscribed, so I can’t give a detailed review yet, but found this to be really interesting. Especially geared toward camping and RVing.

Deep Dive

We posted Marc Andreessen’s article about AI a few weeks ago:

He gave an optimistic view of AI and the future it will bring. But is he right? In an article this month, Wired takes a counter-position:

Marc Andreessen Is (Mostly) Wrong This Time

Andreessen is incredibly optimistic about the future of AI:

In Andreessen’s promised AI-augmented world, “every child will have an AI tutor that is infinitely patient, infinitely compassionate, infinitely knowledgeable, infinitely helpful.” Every adult will have “an AI assistant/coach/mentor/trainer/advisor/therapist” that “will be present through all of life’s opportunities and challenges, maximizing every person’s outcomes.”

But as the article acknowledges, Andreessen is leaving out a few blindspots.

First, the idea that AI will be infinitely knowledgeable. There is certainly the potential for an incredible amount of information, but many of us continue to see the lack of knowledge or understanding in AI. And second, the idea that AI will help us to make better decisions. I’m optimistic that will be the case, but the problem is always about alignment. Better decisions for who?

An “infinitely patient” and “infinitely helpful” AI coach could just as happily help its human master wreak genocide as devise a more efficient manufacturing process or a fairer benefits framework.

While I’m also optimistic that AI will be a productivity multiplier, we have to acknowledge there will be disruption. And we have to figure out how to deal with it in a better way than “learn to code.”

while technology may make companies more productive, it only sometimes makes individual workers more productive (so-called marginal productivity). Other times, it just allows companies to automate part of the work and employ fewer people.

While we may not be facing an AI apocalypse in the near future, we still have a lot to deal with in terms of disruption.

Share

Subscribe now

Here’s this week’s news, product applications, and broader philosophical implications. So here we go:

• 🍿 Pixar

• 🧮 Outsourcing AI Training to AI

• 🛒 App Store

• ⚖️ Regulation

• ✒️ Assistants

• 🎥 Art and Jobs

Latest News and Updates

Pixar Used AI to Stoke Elemental's Flame

According to Pixar, making fire look realistic but not creepy is really difficult. So for the VFX, they are now using neural style transfer (NST) to animate in a way they haven’t done before.

For Sohn, it was an opportunity to make the movie look the way he wanted, while also making something that looked like nothing audiences had ever seen before. It symbolized, he says, one of the things he loves about Pixar: the meeting of art and technology, where the latter is a big part of the process, but only one element.

I haven’t seen the movie yet, but the fact we can use AI to push the boundaries of creativity is pretty exciting. Augmenting what we’ve done (or never done) before.

The people paid to train AI are outsourcing their work… to AI

It was inevitable, but goes to the problem we discussed last week about model collapse. Why would a person continue to do a repetitive, low-value task when we can get AI to do it?

A significant proportion of people paid to train AI models may be themselves outsourcing that work to AI, a new study has found. 

It takes an incredible amount of data to train AI systems to perform specific tasks accurately and reliably. Many companies pay gig workers on platforms like Mechanical Turk to complete tasks that are typically hard to automate, such as solving CAPTCHAs, labeling data and annotating text. This data is then fed into AI models to train them. The workers are poorly paid and are often expected to complete lots of tasks very quickly. 

No wonder some of them may be turning to tools like ChatGPT to maximize their earning potential.

OpenAI Considers Creating an App Store for AI Software

Are you ready for the next wave of chatbots and tools? OpenAI may be launching a marketplace where you can create, sell, and buy customized AI models. It seems like a natural progression.

The company is considering launching a marketplace in which customers could sell AI models they customize for their own needs to other businesses, according to two people with knowledge of discussions at the company.

Biden to meet with A.I. experts in San Francisco

Hot on the heels of the EU moving to regulate AI, the US is starting to think about it?

President Joe Biden will meet with artificial intelligence experts in San Francisco on Tuesday as the U.S. government continues to grapple with how best to regulate the technology, according to a White House official.

The visit comes as many in Washington are seeking to learn more about AI’s risks and benefits for society so they don’t repeat the mistakes around lack of early regulation on social media and other internet technologies.

Google Tells Employees to Stay Away from Its Own Bard Chatbot

I don’t know how to feel about tech giants telling their own employees to stay away from AI chatbot tools. It’s almost like they’re saying to all of us that they know they’re using unethical practices to train their models, and they’re okay with it as long as the data comes from everyone else, but they don’t want their data as part of the models.

Alphabet is reportedly concerned with employees inputting sensitive information into these chatbots since human reviewers may sit on the other end reviewing chat entries. These chatbots may also use previous entries to train themselves, posing another risk of a leak.

Useful Tools & Resources

This week I explored tools to help with AI Meeting Notes. This is a broad topic, so we’ll have more to come. I’ll be creating a library of these tools soon, so remember to check back for everything soon.

Sembly.ai

I began using Sembly.ai for meeting notes. It integrates with my Google calendar, but I can also invite it to meetings from other sources (which is really useful if you’re managing several calendars like I am). I can also use it for in-person meetings through the computer or app. I don’t attend a ton of in-person meetings right now, but the ability to get notes in-person seems really great.

I spoke with the founder of Sembly.ai on my Product by Design podcast and we had a great conversation. You should check it out (and, no, I’m not getting any kickbacks for this, and am paying for my account):

Otter.ai

Otter.ai has been around for a long time, and has some great looking features as well. It is more expensive than Sembly.ai, but looks like it also offers a few more things right now. I will probably take it for a spin in the coming days.

Supernormal AI

Similar to Sembly and Otter, Supernormal offers to attend meetings, make notes, create follow-ups, etc. I’ve just started experimenting with it as well, since it has a free pricing tier.

Meetgeek.ai

I’ll be testing this one soon as well, but Meetgeek has similar functionality to the other offerings here.

Krisp.ai

Krisp also offers similar functionality. I believe it started as a noise reduction app for online meetings, but now has AI assistant functionality.

Deep Dive - AI, Art, and Jobs

Disney received a lot of negative press this week as we learned that AI was responsible for the opening credits of Secret Invasion.

Some outlets reported that some VFX artists lost their jobs on this because of the use of AI. This was walked back by a statement from the studio that worked on the opening:

“The production process was highly collaborative and iterative, with a dedicated focus on this specific application of an AI toolset,” added Method Studios. “It involved a tremendous effort by talented art directors, animators (proficient in both 2D and 3D), artists, and developers, who employed conventional techniques to craft all the other aspects of the project. However, it is crucial to emphasize that while the AI component provided optimal results, AI is just one tool among the array of toolsets our artists used.”

Many artists and creators are rightly concerned about the use of AI in creative endeavors. The Writers Guild is currently on strike in order to codify that AI can’t replace writers but only be used for idea generation.

It certainly doesn’t seem like there is an imminent threat of AI replacing VFX artists or writers, especially given some of what we’ve seen so far. But AI is rapidly improving, and it’s difficult to see what jobs it won’t eventually touch.

It’s going to be a difficult balance. In my work, I’ve used AI to reduce the time it takes to do certain tasks by 10X. As I’ve created documents over the past few weeks, I’ve been able to draft up long documents in a matter of minutes. Then spend an hour or two customizing and editing based on my expertise. In the past, these have take me 10-15 hours, and now I can do them in 2-3.

So AI has become a force multiplier in my work. I’m not a VFX artist, but I understand that going through frame by frame to add FX is a long and difficult process. The ability to multiply what an artist can do through the use of AI seems like an incredible opportunity. But that ultimately depends on how we use it. Does it mean we only need a couple VFX artists, or does it mean we can use all the artists to create much more creative content?

I hope it is the latter. That is what I see for the future in engineering and coding. We can do much more. So rather than cut jobs or people, we can finally create 10X the things that were possible before.

AI isn’t going away. So we’ll need to continue to grapple with these difficult questions. I expect there will be many more missteps like the one Disney had. But we need to have these conversations and face this new reality head on. I’m optimistic we’ll be able to embrace change, multiplying human creativity more than ever before. But it’s going to be a bumpy ride.

Share

Subscribe now

Here’s this week’s news, product applications, and broader philosophical implications. So here we go:

• 💰 $4.4 trillion

• 🧮 Model Collapse

• 🛒 Online Shopping

• 🎵 The Beatles

• ✒️ Prompting

• 🇪🇺 AI Act

Latest News and Updates

The economic potential of generative AI: The next productivity frontier

A new report released by McKinsey this week suggests that AI could add $4.4 trillion dollars of value to the global economy each year going forward. It could automate 70% of some jobs, mostly among knowledge workers.

Generative AI’s impact on productivity could add trillions of dollars in value to the global economy. Our latest research estimates that generative AI could add the equivalent of $2.6 trillion to $4.4 trillion annually across the 63 use cases we analyzed—by comparison, the United Kingdom’s entire GDP in 2021 was $3.1 trillion. This would increase the impact of all artificial intelligence by 15 to 40 percent. This estimate would roughly double if we include the impact of embedding generative AI into software that is currently used for other tasks beyond those use cases.

The AI feedback loop: Researchers warn of ‘model collapse’ as AI trains on AI-generated content

We’re only 6 months into our new age of generative AI, but already we’re seeing what could be a huge problem: model collapse. Most of what AI models have been trained on until now has been human-produced content. But with so much AI content being generated, we’re seeing more and more of that being included in AI model training. And the models don’t do well when trained on generative AI content.

Now, as more people use AI to produce and publish content, an obvious question arises: What happens as AI-generated content proliferates around the internet, and AI models begin to train on it, instead of on primarily human-generated content?

A group of researchers from the UK and Canada have looked into this very problem and recently published a paper on their work in the open access journal arXiv. What they found is worrisome for current generative AI technology and its future: “We find that use of model-generated content in training causes irreversible defects in the resulting models.”

Google’s New AI Tool Is About to Make Online Shopping Even Easier

Imagine having an avatar of yourself that could try on clothes from all sorts of brands and help you decide what will look good and what you want to buy? That’s what Google is creating. Shopping for clothes online has always been more difficult than other goods, so this will be an interesting improvement.

Now, customers in the United States can virtually “try on” women’s tops. The company uses images of real models ranging from XXS to 3XL to wear AI-generated versions of clothes from hundreds of brands sold across Google, like Anthropologie, Everlane, and H&M. You can scroll through and select different body types or skin tones and see how clothes might drape on your own body. When you find the model that most closely resembles you, you can save them to be your default model. 

Paul McCartney says AI tools helped rescue John Lennon vocals for ‘last Beatles record’

It’s not just living artists who will have AI create new music, but any musician.

According to McCartney, technology developed for the recent Beatles documentary Get Back was able to extract former bandmate John Lennon’s vocals from a low-quality cassette recording in order to create the foundation for the track.

Useful Tools & Resources

This week I explored tools to help with AI Prompting. This is a broad topic, so we’ll have more to come. I’ll be creating a library of these tools soon, so remember to check back for everything soon.

PromptPal

PromptPal let’s you search through all different types of prompts, for GPT, Dall-E, Midjourney, etc. It also has specific categories like Product Management prompts, which I’ll have to try more.

But if you are looking for specific ideas on prompts, this will get you a good start.

PromptBase

PromptBase is a site where you can create prompts yourself, buy prompts that others have created, and sell one that have worked well. If you don’t want to take the time to learn prompting, this seems like a good place to get what you need quickly. And if you are good at prompts, a place to see them.

HuggingFace

This one has some interesting potential. It didn’t do what I had envisioned after I used one of the prompts, but this is why we test lots of things, right?

Noonshot

I really liked Noonshot for Midjourney prompting. It gives you lots of guidance for what to add. Again, not at all what I anticipated, but that probably speaks to my need for more time in Midjourney.

Deep Dive - EU Regulation

The European Union has been leading the way (generally) when it comes to regulation and protecting users on the internet and with new technology. And this week has moved forward with significant steps in regulating AI with the AI Act.

According to the NY Times:

The European Parliament, a main legislative branch of the European Union, passed a draft law known as the A.I. Act, which would put new restrictions on what are seen as the technology’s riskiest uses. It would severely curtail uses of facial recognition software, while requiring makers of A.I. systems like the ChatGPT chatbot to disclose more about the data used to create their programs.

The EU has traditionally been at the forefront of thinking about these issues.

The European Union is further along than the United States and other large Western governments in regulating A.I. The 27-nation bloc has debated the topic for more than two years, and the issue took on new urgency after last year’s release of ChatGPT, which intensified concerns about the technology’s potential effects on employment and society.

Being very thoughtful about how to protect people while fostering innovation is a difficult balance. According to the European Parliament:

Parliament’s priority is to make sure that AI systems used in the EU are safe, transparent, traceable, non-discriminatory and environmentally friendly. AI systems should be overseen by people, rather than by automation, to prevent harmful outcomes.

Wisely, the EU is segmenting AI into different categories:

• Unacceptable Risk

• High Risk

• Limited Risk

• Minimal/No Risk

But regulation isn’t easy. First, governments don’t move quickly while technology, and especially AI, has been moving faster than anything we’ve ever seen.

According to the Brookings Institute:

To keep the corporate AI race from becoming reckless requires the establishment and development of rules and the enforcement of legal guardrails. Dealing with the velocity of AI-driven change, however, can outstrip the federal government’s existing expertise and authority. The regulatory statutes and structures available to the government today were built on industrial era assumptions that have already been outpaced by the first decades of the digital platform era. Existing rules are insufficiently agile to deal with the velocity of AI development.

And like the EU has been thinking about, we have to consider what to regulate vs. what not to (or what to regulate differently).

Because AI is a multi-faceted capability, “one-size-fits all” regulation will over-regulate in some instances and under-regulate in others. The use of AI in a video game, for instance, has a different effect—and should be treated differently—from AI that could threaten the security of critical infrastructure or endanger human beings. AI regulation, thus, must be risk-based and targeted.

Additionally, should applications be regulated, or the actual development of AI? According to WSJ:

Tech companies and their lobbyists argue that any government-enforced rules should focus on specific AI applications—and not put too many restrictions on how AI is developed, as is being proposed in Europe. They say such an approach would impede innovation. 

And who gets to regulate? Too often, companies get to regulate themselves. Sometimes that works, other times it doesn’t. And while the US is traditionally in the lead when it comes to creating technology, the EU has typically created the frameworks for regulation, especially since it moves earlier and more comprehensively.

Once again it appears as though the EU, which has been in the lead in establishing digital platform policy with its Digital Markets Act and Digital Services Act, is also in the lead on establishing AI policy. On June 14 the European Parliament overwhelmingly approved the AI Act. Following its adoption, the regulatory machinery of the European Commission will begin developing enforceable policies.

We certainly need to think carefully about the use of AI and its regulation. How do we continue to benefit from advances in AI while managing the risks and impacts to humans. This is just the beginning.

Share

Subscribe now

Here’s this week’s news, product applications, and philosophical implications. So here we go:

• 🧑‍🤝‍🧑 AI Will Save Us

• 🧮 Algorithms

• 🚐 Meta

• 🤳 Photo Apps

• 🫱‍🫲 Meetings and AI

Latest News and Updates

Why AI Will Save the World

We talk a lot about the potential of AI. The amazing productive potential but also the incredible destructive potential. If you’re looking for a good dose of optimism, Marc Andreessen has you covered:

What AI offers us is the opportunity to profoundly augment human intelligence to make all of these outcomes of intelligence – and many others, from the creation of new medicines to ways to solve climate change to technologies to reach the stars – much, much better from here.

New algorithms will transform the foundations of computing

We know AI will transform many aspects of our technology and our lives. And making things faster and better while using less power is a great start:

AlphaDev uncovered a faster algorithm for sorting, a method for ordering data. Billions of people use these algorithms everyday without realising it. They underpin everything from ranking online search results and social posts to how data is processed on computers and phones. Generating better algorithms using AI will transform how we program computers and impact all aspects of our increasingly digital society. 

Meta plans to put AI everywhere on its platforms

Zuckerberg announced plans to put AI into everything Facebook is doing.

Meta CEO Mark Zuckerberg announced a plan to employees on Thursday that will see it put generative AI text, image and video generators into its flagship products, such as Facebook and Instagram.

Facebook's parent company has been at the forefront of generative AI research, but has been somewhat slow to put those technologies into its products.

Instagram is apparently testing an AI chatbot that lets you choose from 30 personalities

In line with the announcement from Meta, we got more details on Instagram’s approach from a leak this week. Following Snap’s lead, Instagram may be getting ready to launch its own AI tools:

A screenshot shared by leaker Alessandro Paluzzi on Twitter shows what seems to be an intro screen for the new Instagram feature. It says the chatbots will be able to answer questions, give advice, and help users write messages. It also says that users will be able to choose between “30 AI personalities and find which one you like best.”

Charlie Brooker Got ChatGPT To Write A Black Mirror Episode, But It Was ‘Shit’

Are we at risk of having all our content be AI produced? It doesn’t seem like it yet. While ChatGPT and other tools can write pretty good documents and excel in certain areas, they aren’t likely to take over real creative endeavors soon.

“I’ve toyed around with ChatGPT a bit," Brooker reveals in the new issue of Empire. "The first thing I did was type 'generate Black Mirror episode' and it comes up with something that, at first glance, reads plausibly, but on second glance, is shit. Because all it’s done is look up all the synopses of Black Mirror episodes, and sort of mush them together. Then if you dig a bit more deeply you go, 'Oh, there’s not actually any real original thought here.' It’s [1970s impressionist] Mike Yarwood — there’s a topical reference.”

This makes sense, especially when we understand how LLMs work. But a Black Mirror episode written by ChatGPT would be on brand for sure. If it was anything close to passable.

Useful Tools & Resources

This week I explored tools to help with photos. This is a broad topic, so we’ll have more to come. I’ll be creating a library of these tools soon, so remember to check back for everything soon.

We are about to have a high school reunion, so that has meant many of us from our class are busy scanning photos. Which led me to naturally look at some of the tools available. I haven’t had a chance to use all of them, but let me highlight a few things.

Reimagine

I took this photo from our yearbook and scanned it in using the Reimagine AI app:

I then colorized it and enhanced it:

It definitely sharpens the image. Unfortunately, it didn’t get the color of the shirts right. They were bright yellow. And some of the faces, on close inspection, aren’t quite right.

They look good from afar, and are recognizable, but are just different enough that it feels a little strange. Especially knowing exactly how everyone should look.

Overall, though, I thought the app did a good job. It has some things that need to be worked out, but it’s pretty solid, especially comparing to the Photoscan app (from Google) that is generally my go-to.

Deep Dive

I had the opportunity to sit down with Artem Koren and discuss a topic that I think is incredibly pertinent and fascinating: AI and meetings.

Artem is the Chief Product Officer and Co-Founder of Semby AI, a SaaS platform that uses proprietary AI algorithms to transcribe and analyze meetings. Artem has 16+ years of product management experience and has applied AI learning to his work since 2009.

We had a fascinating discussion about Sembly AI and meetings in general. As Artem put it, "technology has been facilitating meetings in the workspace, but not really understanding what they're about--the content of that."

Meetings are part of the normal workday for all of us. But they are almost always less than optimal. I explored this in a post a while back:

Product Thinking

(De)Prioritizing Meetings

It’s 5:00 on a weekday. You stand up to stretch. You’ve been in meetings (or on Zoom calls) all day with barely a break. You’ve been multitasking as much as you can, shooting out emails and Slack messages whenever you can divert your attention, coming back to the calls whenever you need to chime in or listen or catch up on something you missed…

Read more

6 years ago · Kyle Evans

But AI presents a new possibility for solving the problems with meetings. From Tech Brew:

As companies grapple with how to put generative AI tech to work in the office, taking notes on meetings is emerging as a clear use case.

Salesforce said last month it would roll out a suite of tools including a summarization feature for calls and messages in Slack; Zoom recently teamed with OpenAI and Anthropic to offer similar tools for video meetings; and Microsoft now includes AI-generated recaps in its subscription tier for Teams.

Sembly AI has also been doing this for years, along with some other cool things.

One of the most interesting things for me, is how does the ability to have an AI agent for meetings or conversations or other scenarios take us? If your AI agent can attend meetings for you soon, what’s next?

Beyond meeting summaries, Patel sees generative AI as a major shift in the technology landscape that will ultimately transform the way people interact with computers in the workplace and beyond.

We discussed these ideas on the podcast. Why wouldn’t you be able to begin to have AI only meetings? And what comes after that?

It’s a fascinating space, and we’ve got more coming very soon. You’ll want to listen to the whole conversation and stay tuned for several more that dive into these areas.

Share

Subscribe now

Here’s this week’s news, product applications, and philosophical implications. So here we go:

• 🦴 Human Extinction

• 💭 Hallucinations

• 🇯🇵 Japan Goes All In

• 📄 PDF Tools

• ↔️ The Alignment Problem

Latest News and Updates

Experts are warning AI could lead to human extinction. Are we taking it seriously enough?

This week, we saw many people again voice the concern of major repercussions to AI:

On Tuesday, hundreds of top AI scientists, researchers, and others — including OpenAI chief executive Sam Altman and Google DeepMind chief executive Demis Hassabis — again voiced deep concern for the future of humanity, signing a one-sentence open letter to the public that aimed to put the risks the rapidly advancing technology carries with it in unmistakable terms.

“Mitigating the risk of extinction from AI should be a global priority alongside other societal-scale risks such as pandemics and nuclear war,”

I posted previously about how we are facing a moment similar to the creation of atomic weapons, with all the destructive capabilities. As humans, we nearly took ourselves to the brink with that one:

@ai.nexthorizondangers of AI and the future of the technology#chatgpt #ai #greenscreen

Tiktok failed to load.

Enable 3rd party cookies or use another browser

OpenAI is pursuing a new way to fight A.I. ‘hallucinations’

I was generating a document this week, and after feeding in information about my background, ChatGPT generated a few paragraphs about me. Surprisingly, it included the phrases “despite not having an MBA” and “having 5 years of product experience.” I do, in fact, have an MBA and have 15 years of product experience. It was strange to see ChatGPT generate false information about me, when I literally had input correct information moments before.

This problem isn’t unusual. And OpenAI is working on ways to fight it:

AI hallucinations occur when models like OpenAI’s ChatGPT or Google’s Bard fabricate information entirely, behaving as if they are spouting facts. One example: In Google’s own February promotional video for Bard, the chatbot makes an untrue claim about the James Webb Space Telescope.

Japan Goes All In: Copyright Doesn’t Apply To AI Training

Japan has decided to go all in for AI.

In a surprising move, Japan’s government recently reaffirmed that it will not enforce copyrights on data used in AI training. The policy allows AI to use any data “regardless of whether it is for non-profit or commercial purposes, whether it is an act other than reproduction, or whether it is content obtained from illegal sites or otherwise.”

We’ll see if this sets a precedent for other countries as well. It certainly has the potential to unleash a lot of generative AI training.

It seems the Japanese government believes copyright worries, particularly those linked to anime and other visual media, have held back the nation’s progress in AI technology. In response, Japan is going all-in, opting for a no-copyright approach to remain competitive.

A majority of Americans have heard of ChatGPT, but few have tried it themselves

It’s easy to think that everyone has heard of ChatGPT, AI tools, etc. And that everyone is using them. But that’s just not the case. When we’re surrounded by these tools, and other people who are tech savvy, it’s easy to think everyone knows about them. We forget that most people don’t adopt technology as quickly as those of us at the forefront. According to Pew Research:

About six-in-ten U.S. adults (58%) are familiar with ChatGPT, though relatively few have tried it themselves, according to a Pew Research Center survey conducted in March. Among those who have tried ChatGPT, a majority report it has been at least somewhat useful.

However, few U.S. adults have themselves used ChatGPT for any purpose. Just 14% of all U.S. adults say they have used it for entertainment, to learn something new, or for their work. This lack of uptake is in line with a Pew Research Center survey from 2021 that found that Americans were more likely to express concerns than excitement about increased use of artificial intelligence in daily life.

Useful Tools & Resources

This week I explored tools to help summarize and analyze PDFs. I’ll be creating a library of these tools soon, so remember to check back for everything soon.

ChatPDF

This was one of my favorite tools. It analyzed the documents I uploaded and then prompted me with questions I could ask:

@ai.nexthorizonchatPDF - using AI PDF tools to summarize and analyze PDF docs #ai

Tiktok failed to load.

Enable 3rd party cookies or use another browser

PDF.AI

This was another really useful tool as well. I uploaded a research paper (the same as above) and was able to ask questions.

@ai.nexthorizonpdf.ai - taking a look at some of the AI tools for pdfs #ai

Tiktok failed to load.

Enable 3rd party cookies or use another browser

ChatWithPDF

There are a number of other tools as well, including plugins for ChatGPT like ChatWithPDF. Unfortunately when I used this plugin, it very confidently gave me a lot of incorrect information for the PDF. So, as with everything, you need to have some knowledge of the subject to check the work.

Deep Dive - The Alignment Problem

I dove into the book The Alignment Problem by Brian Christian this week. If you missed it, definitely check out the podcast above and the post here.

The heart of the alignment problem is ensuring that our machines—the AI we create, the algorithms and models—understands our objectives and values. That it understands the intent behind the instructions, rather than just the text it may see.

As Quanta magazine puts it:

Computers frequently misconstrue what we want them to do, with unexpected and often amusing results. One machine learning researcher, for example, while investigating an image classification program’s suspiciously good results, discovered that it was basing classifications not on the image itself, but on how long it took to access the image file — the images from different classes were stored in databases with slightly different access times. Another enterprising programmer wanted his Roomba vacuum cleaner to stop bumping into furniture, so he connected the Roomba to a neural network that rewarded speed but punished the Roomba when the front bumper collided with something. The machine accommodated these objectives by always driving backward.

These examples are amusing. We’ve seen similar problems in video games when we put AI to work. We may give points for not dying, which leads to the character never moving. Or give points for possessing the ball in a soccer game, which leads to lots of possession and never to scoring goals.

What happens when there is more at stake?

But the community of AI alignment researchers sees a darker side to these anecdotes. In fact, they believe that the machines’ inability to discern what we really want them to do is an existential risk. To solve this problem, they believe, we must find ways to align AI systems with human preferences, goals and values.

Of course, it’s not enough to understand goals:

If you believe that intelligence is defined by the ability to achieve goals, that any goal could be “inserted” by humans into a superintelligent AI agent, and that such an agent would use its superintelligence to do anything to achieve that goal, then you will arrive at the same conclusion that Russell did: “All that is needed to assure catastrophe is a highly competent machine combined with humans who have an imperfect ability to specify human preferences completely and correctly.”

This is the classic paper clip problem many of us are familiar with. As New Scientist puts it:

This goes back to 2003, when Nick Bostrom, a philosopher at the University of Oxford, posed a thought experiment. Imagine a superintelligent AI has been set the goal of producing as many paper clips as possible. Bostrom suggested it could quickly decide that killing all humans was pivotal to its mission, both because they might switch it off and because they are full of atoms that could be converted into more paper clips.

The way to solve for this won’t be easy, but will require everyone working together and addressing the problem. As Paul Christiano, a researcher at OpenAI, put it on Effective Altruism:

“Rather than taking our initial teacher to be one human and learning from that one human to behave as well as a human, let’s start by learning from a group of humans to do more sophisticated stuff than one human could do.” That could give us an AI, which isn't just “human-level,” but is the level of a human group.

Subscribe now

Share

I’m iterating on the format, but like the idea of news from the week, exploring specific tools and products, and diving into deeper implications of AI. News, product, philosophy (sums me up as well). So here we go:

• 👩‍✈️ Microsoft Copilot

• 🦾 Humanoid Robots

• 🖼️ Photoshop

• 🖥️ Presentation Tools

• 🧘 Who Are We After AI?

Subscribe now

Latest News and Updates

The 5 biggest announcements from Microsoft Build 2023

Microsoft came out big with Build 2023. It was AI-focused, as you can imagine. And they announced their AI copilot for Windows 11 and Office 365, and Bing for ChatGPT.

The company has made several announcements on how it’s expanding its use of AI across its apps and services, including in Windows 11, Microsoft 365, and more.

AI startup Figure raises $70 million to build humanoid robots

This was inevitable, and will likely accelerate.

The new funding will be used to accelerate the development and manufacturing of the company’s first autonomous humanoid, which is expected to launch in the next few months.

"We believe general-purpose humanoid robots have far more potential than single-purpose robots," said Adcock, "The deployment into the workforce can help address labor shortages and over time lead the way in eliminating the need for unsafe and undesirable jobs."

Anthropic Raises $450 Million in Series C Funding to Scale Reliable AI Products

No signs of slowing down in the AI space, and funding continues to grow among all the different companies. Anthropic, the builder of Claude, has raised $450 million this week.

We are pleased to announce that we have raised $450 million in Series C funding led by Spark Capital with participation from Google, Salesforce Ventures, Sound Ventures, Zoom Ventures, and others. The funding will support our continued work developing helpful, harmless, and honest AI systems—including Claude, an AI assistant that can perform a wide variety of conversational and text processing tasks.

Photoshop has entered the chat

Photoshop is synonymous with photo editing and manipulation. So it isn’t surprising that Adobe has added AI generation to the popular editing tool.

The new AI tools will work similarly to other AI image creation technology, such as Midjourney and DALL-E, which allow people to produce pictures from simple text prompts.

Nvidia to the Moon

AI is taking a lot of companies to record highs. That includes chip maker Nvidia.

Nvidia has come to dominate the market for chips used in artificial intelligence (AI) systems.

But all that would not be possible without powerful computer hardware - in particular computer chips from California-based Nvidia.

Originally known for making the type of computer chips that process graphics, particularly for computer games, Nvidia hardware underpins most AI applications today.

Useful Tools & Resources

This week I explored tools to help create slides and presentations. I’ll be creating a library of these tools soon, so remember to check back for everything soon.

SlidesAI.io

This tool integrates directly with Google Slides. I took an article I wrote for Product Thinking, pasted it in, and asked for a presentation. And got one. So if you have written content (less than 2500 words for the free version), you can easily generate a presentation.

@ai.nexthorizonGenerating AI slides - stop making your own presentations #slides#chatgpt #ai

Tiktok failed to load.

Enable 3rd party cookies or use another browser

Tome.app

Tome took my prompt to create a presentation on “product roadmaps” and ran with it. It looks really good as well. I’d need to edit quite a bit if I wanted it to align with my article, but the fact I could simply create a presentation in a minute on a topic is incredible.

Sendsteps

I was impressed by the content generated by my simple prompt on this one. It gives more of a framework and would need more editing, but also gives lots of presenter notes. Not my first choice in graphics for the subject, but I’d play around with that some more to fine tune it.

Chatba.com

I wasn’t able to create a product roadmap presentation with Chatba since traffic was too high, so I had to settle for a cached prompt about the failure of Zune. It seems simpler than the other presentations both in style and content, which may work for you.

Deep Dive

Who will we be when AI takes our jobs? Or at least disrupts significant portions of them?

Job disruptions aren’t unusual. As new technology advances, it always disrupts the old way of doing things. According to The Atlantic:

“Before, progress was linear and predictable. You figured out the steps and the computer followed them. It followed the procedure; it didn’t learn and it didn’t improvise,” the MIT professor David Autor, one of the world’s foremost experts on employment and technological change, told me. ChatGPT and the like do improvise, promising to destabilize a lot of white-collar work, regardless of whether they eliminate jobs or not.

It is difficult to say how many jobs will be disrupted, especially among the highly educated. But the rapid advance of AI may make this disruption unusual, according to Wired:

we've never had a wave of automation in which white-collar workers are uniquely vulnerable, and we should therefore expect this one to play out differently.

The crux of the difference lies in the relationship that blue-collar and white-collar workers have with work. According to one study, white-collar workers tend to feel that they “express” their “full potential” more at work than blue-collar workers do; they also experience higher levels of “developing inner self” at work. 

While many who work in manual labor (rightly) identify their work as a means to an end—a paycheck—many of us who work in white-collar jobs (wrongly) develop our identities around our jobs.

This is due to a number of factors. For many of us, our work becomes an art form. We view it as a creative expression or ourselves, and so to see a machine be able to master the appearance of creativity is shocking at first.

At the furthest edge of white-collar work is a species of task where competency is so admired that it becomes a sport or an art, and competency is rewarded by society with status and esteem, over and above financial compensation. This is the category of games of logic and art. Our shock at the new wave of AI models like ChatGPT and Midjourney comes from their proficiency at the artier, more creative tasks like writing and illustration. But the more strictly logical sports of chess and Go have long since been conquered by previous waves of AI, and so it may be instructive for the left-brainers to look in on how the right-brainers have handled usurption, emotionally and practically. 

How have other areas handled this type of disruption? In Go, the top players took different paths. For some, it pushed them out of the game and into depression. For others, they joined AI team (if you can’t beat them, join them). And still for others, they shifted from their old approach to Go to a new approach, and have thrived.

Important to this story, too, is that, unlike Fan, whose pivot to temporary AI research consultant could be seen as a demotion from European Go champion, Ke’s pivot allowed him to remain at the top of the game.

We’re likely to see a massive shift in the coming years based on the technology we’ve created. And it is already happening according to the LA Times, slowly, then rapidly:

The maddening thing is that, at least at first, it’s probably not going to look like much of anything. There will be no cybernetic android that lumbers over to your desk and takes over your work duties, no disembodied robot voice that suddenly assumes command over your department. It might just look like routine layoffs, or a freelancer having more trouble finding work.

If you’ve not seen this yet, you probably will.

In the end, we’re still getting more AI and fewer humans — even if the AI is not doing the humans’ job, exactly. AI will be used by executives and managers much this way, I think: to help gin up investment in future-forward products that require less labor costs to make, to buffer layoffs or attrition in human departments, and to hire more part-time or project-based workers.

The question isn’t if it happens, but more about how and when it does. And how we react individually and collectively. Can you individually make the shift to a new world? And for those who struggle, what can we collectively do to ease the transition. We’ve historically been very bad about this. Not everyone can just “learn to code” or learn to AI.

Hopefully we can do better this time around. And if we can’t, maybe AI can help us.

Subscribe now

Share

• 🔌 ChatGPT Plugins

•  Apple Bans ChatGPT

• 🎤 Sam Altman testifies

• 🖥️ Zoom gets AI

• 🙈 The Cost of AI Training

Subscribe now

Latest Updates

ChatGPT Plugins

OpenAI announced it would begin rolling out access to plugins to premium users this week.

If you are a ChatGPT Plus user, enjoy early access to experimental new features, which may change during development. We’ll be making these features accessible via a new beta panel in your settings, which is rolling out to all Plus users over the course of the next week.

@ai.nexthorizonplugins for chatGPT #ai #chatgpt

Tiktok failed to load.

Enable 3rd party cookies or use another browser

I got access on Monday and began playing around with a number of them. If you are a ChatGPT Plus user, you should have access as well.

Apple Bans ChatGPT

It may not come as a surprise, since Microsoft is a major competitor to Apple, but Apple doesn’t want their employees using ChatGPT.

@ai.nexthorizonApple bans chatGPT for it's employees #chatgpt #apple #ai

Tiktok failed to load.

Enable 3rd party cookies or use another browser

Apple employees will reportedly be restricted from using ChatGPT and other artificial intelligence tools.

This follows in the footsteps of other companies blocking the use of AI tools.

Apple isn't the first major tech company to restrict ChatGPT use by its staff. ChatGPT has been banned for Samsung employees after sensitive information was inadvertently leaked to the platform. Other institutions, like JPMorgan, Bank of America, and Citigroup have also banned ChatGPT to protect confidential information.

It will certainly be interesting to see the divides between people and companies as the AI wars continue to heat up.

Sam Altman Testifies

Sam Altman, CEO of OpenAI, testified before congress this week. He urged regulation in the industry:

“We think that regulatory intervention by governments will be critical to mitigate the risks of increasingly powerful models”

It’s not clear to me that the US government understands technology, let alone AI and the speed it is moving. And regulation never moves quickly. So this will be an interesting space to watch.

Anthropic and Zoom

Zoom and Anthropic announced a new partnership. Zoom will begin to incorporate Claude, Anthropic’s AI into Zoom tools:

Zoom will use Claude, our AI assistant built with Constitutional AI, to build customer-facing AI products focused on reliability, productivity, and safety.

Lots more to come, no doubt.

Other Links

Clippy Office Assistant - Do you need a rude AI assistant? This should fit the bill.

Product Photos - Get better product photos with AI. I’m trying this out over the weekend to update some listings.

Stratup.AI - Need help generating a startup idea?

Chatbot Creator - Platform for creating your own chatbot.

Midjourney Prompt Generator - Need help generating prompts? The image below came from the prompt generator I used to help me depict the next deep dive

Deep Dives

In a funny scenario in Silicon Valley, Erlich and Jian-Yang create an app that detects “hot dog” or “not hot dog.”

Of course, this isn’t a very useful tool for food, but has useful implications for detecting anatomic imagery, and Periscope purchases the app for $4 million dollars. But that leaves Dinesh to train the model on what is a hot dog and what isn’t:

Gilfoyle finds this hilarious, naturally. And it is funny for the characters in the show.

But the reality is that content moderation and AI training has a dark side. We often overlook the people who have to train the models to identify explicit content in order to help the rest of us avoid it.

In an interesting article, workers in Kenya were employed to train ChatGPT to identify and label explicit content. That meant reading the content and labeling it for ChatGPT to understand.

In a process called Reinforcement Learning from Human Feedback, or RLHF, bots become smarter as humans label content, teaching them how to optimize based on that feedback. AI leaders, including OpenAI’s Sam Altman, have praised the practice’s technical effectiveness, yet they rarely talk about the cost some humans pay to align the AI systems with our values.

This can take a heavy toll on the workers who are doing the moderation and training:

Mophat Okinyi, a QA analyst on Mathenge’s team, is still dealing with the fallout. The repeated exposure to explicit text, he said, led to insomnia, anxiety, depression, and panic attacks. Okinyi’s wife saw him change, he said, and she left him last year. “However much I feel good seeing ChatGPT become famous and being used by many people globally,” Okinyi said, “making it safe destroyed my family. It destroyed my mental health. As we speak, I’m still struggling with trauma.”

This isn’t a new problem, but it is still a serious one. In an article in The Verge, they discuss how Facebook handles its content moderation by asking its (mostly contractors) to review the most obscene content and decide if it violates standards:

Here is a racist joke. Here is a man having sex with a farm animal. Here is a graphic video of murder recorded by a drug cartel. Some of the posts Miguel reviews are on Facebook, where he says bullying and hate speech are more common; others are on Instagram, where users can post under pseudonyms, and tend to share more violence, nudity, and sexual activity.

Each post presents Miguel with two separate but related tests. First, he must determine whether a post violates the community standards. Then, he must select the correct reason why it violates the standards. If he accurately recognizes that a post should be removed, but selects the “wrong” reason, this will count against his accuracy score.

All of us benefit from the work that contractors and employees do behind the scenes to identify harmful content. But it’s not without significant cost. And that’s important to remember. Because we can’t continue to use up real people to clean up the messes society makes with technology.

Subscribe now

Share

• 🧠 Memory Upgrades

• 🏰 Google has a moat now?

• 🖨️ Box

• 🎵 AI Music

Subscribe now

Latest Updates

Anthropic’s memory goes from goldfish to elephant

“Anthropic expanded the context window for Claude — its flagship text-generating AI model, still in preview — from 9,000 tokens to 100,000 tokens. Context window refers to the text the model considers before generating additional text, while tokens represent raw text.”

Why does this matter?

“models with small context windows tend to “forget” the content of even very recent conversations — leading them to veer off topic. After a few thousand words or so, they also forget their initial instructions.”

Google I/O

Google made a ton of announcements this week about their AI work. If you want to sign up to experiment, you can do that as well.

It’s clear Google is wasting no time accelerating everything AI related. And they are in prime position to do it. So much of what we use in our personal and professional lives is a Google product (for many of us). So infusing AI into those products rather than allowing other AI companies to dominate the space seems like a smart move. And infusing they are.

Also, if you haven’t had a chance to try Bard, it is now available for everyone. So you should try it out.

@ai.nexthorizonchatGPT vs Google Bard: User story prompt #chatgpt #bard #ai

Tiktok failed to load.

Enable 3rd party cookies or use another browser

Box CEO Aaron Levie Breaks Down Box AI and Generative AI’s Impact on Business

A good podcast discussion on Box’s approach to AI and some of the considerations they are making as they incorporate AI into their business on behalf of other businesses. This is something I’m thinking a lot about, and probably many of you are as well.

AI Generated Music

AI generated music continues to make the news, with Spotify removing thousands of songs from Boomy. But that represents only a fraction of the songs Boomy has on Spotify. So AI generated music is here to stay and will likely continue to grow like crazy…

@ai.nexthorizonai generated music is taking over rapidly #ai #music #chatgpt #greenscreen

Tiktok failed to load.

Enable 3rd party cookies or use another browser

Not Just the Meta Verse

Meta announces generative AI features for advertisers:

“Meta today announced an AI Sandbox for advertisers to help them create alternative copies, background generation through text prompts and image cropping for Facebook or Instagram ads.”

Wendy’s tests an AI chatbot that takes your drive-thru order

It’s a natural evolution, and I’ve been impressed at a few drive-thrus when I’ve used bots before. So we’ll see how this tech continues to mature.

“The fast-food chain is partnering with Google to train an AI model that knows what ‘biggie bags’ and ‘JBC’ mean.”

Other Links

Recap - Summarize pages with ChatGPT

Botsonic - Create a ChatGPT chatbot and train with your own data

Candide AI - Help you kids learn AI (something I’ve been keenly looking into across a number of tools and sites)

Custom AI Bots - A guide to AI chatbots

Share

• 🧠 AI is decoding brain signals

• 🏰 Google has no moat

• 🖨️ Replacing Jobs

• ⌨️ StarCoder release

• 👴 Godfather of AI

Subscribe now

Latest Updates

Mind-Reading AI

The idea of “reading minds” is infinitely complex, given that our thoughts are not pages in a book. But we seem to be making rapid progress in deciphering what people are thinking: “In a study published in the journal Nature Neuroscience, the researchers described an A.I. that could translate the private thoughts of human subjects by analyzing fMRI scans.”

Google Has No Moat

In a leaked paper from Google, one of their researchers admits they have no moat, but neither does any other AI company: “We have no secret sauce. Our best hope is to learn from and collaborate with what others are doing outside Google. We should prioritize enabling 3P integrations.”

IBM Plans to Replace 7,800 Jobs with AI

It’s hard to tell if AI is simply a scapegoat for layoffs at this point, or if it is beginning to have on impact on hiring. I expect the former, and the latter will come eventually. “IBM Chief Executive Officer Arvind Krishna has revealed plans to pause hiring for about 7,800 positions that could be replaced by artificial intelligence systems over time, according to a Bloomberg news report published Monday.”

The World Economic Forum also projected that AI will be responsible for layoffs over the next 5 years: “Artificial Intelligence and other economic drivers will result in 83 million job losses over the next five years, amid “structural labour market churn” according to the World Economic Forum’s new Future of Jobs report.”

@ai.nexthorizonjobs, layoffs and AI #chatgpt #ai #work #greenscreen

Tiktok failed to load.

Enable 3rd party cookies or use another browser

StarCoder: A State-of-the-Art LLM for Code

Hugging Face and ServiceNow released their open source answer to GitHub’s Copilot: StarCoder. “StarCoder and StarCoderBase are Large Language Models for Code (Code LLMs) trained on permissively licensed data from GitHub, including from 80+ programming languages, Git commits, GitHub issues, and Jupyter notebooks.”

Other Links

KidGeni - A pretty cool generative image tool specifically for kids. You can even order shirts based on the best things your kids create. Or just let them be creative, which is pretty neat.

Chatg - Get information about companies in industry segments. I’ve been consulting for a company regarding software for internal tools, and I’ve experimented with this (along with other tools) to inform some of our decision-making.

Notion - Send tasks to Notion with your voice. I use Notion for everything, so I’m excited to try this out.

SlackGPT - Slack is bringing AI into their tools in a big way.

AudioPen - A way to convert messy thoughts into clear text. I tend to ramble when I speak, so this has such amazing potential, I’m excited to try it out.

Deep Dive

The Godfather of AI Leaves Google and Warns of Danger Ahead

Geoffrey Hinton, a long-time AI researcher, left Google this week so he could speak more freely about his concerns about AI without hurting Google.

@ai.nexthorizondangers of AI and the future of the technology#chatgpt #ai #greenscreen

Tiktok failed to load.

Enable 3rd party cookies or use another browser

“But gnawing at many industry insiders is a fear that they are releasing something dangerous into the wild. Generative A.I. can already be a tool for misinformation. Soon, it could be a risk to jobs. Somewhere down the line, tech’s biggest worriers say, it could be a risk to humanity.

‘It is hard to see how you can prevent the bad actors from using it for bad things,’ Dr. Hinton said.”

In another interview, he discussed more of his concerns:

“I have suddenly switched my views on whether these things are going to be more intelligent than us,” he said in an interview with MIT Technology Review. “I think they’re very close to it now and they will be much more intelligent than us in the future.... How do we survive that?”

A big part of the problem is the unknown. We simply don’t know what AI will be capable of. Right now we are able to harness the power of AI for incredible things, but we simply don’t know what happens when the AI we’ve trained becomes smarter than we are or expands outside the guardrails we create. What happens when technology begins to manipulate or kill on its own? Or bad actors team up with incredibly intelligent AI to take down governments, societies, or humanity?

And the time table is also unknown. In an article from The Guardian:

“I thought it would happen eventually, but we had plenty of time: 30 to 50 years. I don’t think that any more. And I don’t know any examples of more intelligent things being controlled by less intelligent things – at least, not since Biden got elected.

“You need to imagine something more intelligent than us by the same difference that we’re more intelligent than a frog. And it’s going to learn from the web, it’s going to have read every single book that’s ever been written on how to manipulate people, and also seen it in practice.”

He now thinks the crunch time will come in the next five to 20 years, he says. “But I wouldn’t rule out a year or two. And I still wouldn’t rule out 100 years – it’s just that my confidence that this wasn’t coming for quite a while has been shaken by the realisation that biological intelligence and digital intelligence are very different, and digital intelligence is probably much better.”

Of course, we simply don’t know:

There’s still hope, of sorts, that AI’s potential could prove to be over-stated. “I’ve got huge uncertainty at present. It is possible that large language models,” the technology that underpins systems such as ChatGPT, “having consumed all the documents on the web, won’t be able to go much further unless they can get access to all our private data as well. I don’t want to rule things like that out – I think people who are confident in this situation are crazy.”

So now certainly seems like the time to get a handle on this while all of the technology is in its infancy. But will we be able to?

Share

It’s been a wild week in AI, so let’s catch up on a few things:

• AI and music

• How generative AI works

• Latest releases

Subscribe now

Latest Happenings

The music industry is facing another Napster moment, and it’s going to be AI driven.

@ai.nexthorizonai and music - it's only going to get crazier from here #ai #chatgpt #greenscreen

Tiktok failed to load.

Enable 3rd party cookies or use another browser

Music fans responded with disbelief this week to the release on streaming and social media platforms of the viral song "Heart on My Sleeve."

That’s from an NPR article on Friday. “Heart on My Sleeve” was written by a Ghostwriter, who then used AI to make it sound like a collaboration by Drake and The Weeknd.

Millions of views and listens across platforms stacked up, until Universal Music Group shut it down.

Much like Napster in the early 2000s, this has the potential to seriously disrupt music. For those of us who were downloading music at that time, there was no other way to get digital music, except through illegal file sharing online. So that’s what we did. And the music industry fought it until it was forced to embrace it.

So what will be the future of AI generated music?

It is impossible to tell at this point. But obviously there is no putting this genie back in the bottle. While some artists are opposed to the idea (which there is a great case to be made against AI generated music), others are embracing the idea.

Grimes has welcomed musicians to create new songs with her voice using Artificial Intelligence, saying she would split 50% of royalties on any successful AI-generated track that included her voice.

We’ll see much more in this space, like all others, in the near future. So stay tuned. Until then here is Ariana Grande covering Taylor Swift:

@songsupdzaRespondendo a @e.lannn cardigan – ariana grande (ai)

Tiktok failed to load.

Enable 3rd party cookies or use another browser

Deep Dive and Application

We’ve been hearing non-stop about ChatGPT and generative AI lately. But at its core, what is it and how can we understand it better?

In a great article in the New Yorker, Cal Newport, a Georgetown computer science professor (who is also the author of one of my favorite books: Deep Work), describes the workings of ChatGPT and generative AI in simple terms. The whole article is worth a read, but we’ll pull out a few ideas here.

First, ChatGPT and generative AI models are compiling vast amounts of information and using that to generate the next word, based on query parameters. In a simple form, we can imagine a chat program searching through all its sources and then choosing the most likely word to follow whatever word or phrase it has last generated:

In designing our hypothetical chat program, we will use the same general approach of producing our responses one word at a time, by searching in our source text for groups of words that match the end of the sentence we’re currently writing.

We can make these models better by giving them lots of data to search against and allowing them to generate their own rules, rather than dictating every rule for how to generate the text.

From there, the size and complexity of the neural networks and calculations can get massive, but the foundation of the data generation stays the same. It relies on a vast amount of language processing to combine and recombine text in a way that makes sense.

For many applications, the results can be uncanny. But for others, they can be obviously wrong.

When interacting with these systems, it doesn’t take long to stumble into a conversation that gives you goosebumps. Maybe you’re caught off guard by a moment of uncanny humanity, or left awestruck by the sophistication of a response. Now that we understand how these feats are actually performed, however, we can temper these perceptions. A system like ChatGPT doesn’t create, it imitates. When you send it a request to write a Biblical verse about removing a sandwich from a VCR, it doesn’t form an original idea about this conundrum; it instead copies, manipulates, and pastes together text that already exists, originally written by human intelligences, to produce something that sounds like how a real person would talk about these topics. This is why, if you read the Biblical-VCR case study carefully, you’ll soon realize that the advice given, though impressive in style, doesn’t actually solve the original problem very well.

I also encountered this recently. As I was asking ChatGPT to summarize an example from a book, it left out an important detail. So I asked it to try again. Then when I asked it for a date for the example, it recanted the whole conversation and said the example didn’t exist and it was mistaken. Of course, I knew the example was real because I read the book. But it seemed to doubt itself because I was pushing for additional information it didn’t know.

@ai.nexthorizon @ai.nexthorizon ♬ original sound - ai.nexthorizon

Tiktok failed to load.

Enable 3rd party cookies or use another browser

So what does all of this mean for our work? According to this article:

ChatGPT won’t replace doctors, but it might make their jobs easier by automatically generating patient notes from electronic medical-record entries. ChatGPT cannot write publishable articles from scratch, but it might provide journalists with summaries of relevant information, collected into a useful format.

I have many examples where I believe this will play out as well. There still needs to be human interaction and verification. We’re still in the early stages, but we’ll continue to explore how this will affect all of us.

Additional Links

OpenAI rolls out 'incognito mode' on ChatGPT - If you’re concerned about ChatGPT using your data for its training (a real concern for businesses), you’re in luck. They just rolled out a way to turn off sharing.

Apple to Expand Health Initiatives With AI Health Coach, Report Says - Apple is moving more into AI, specifically within health. If you’re question is “what about Siri,” you’re in good company. It sounds like there is a lot of internal politics at play.

Replit Raises Money, Launches Lots of New Features - The coding copilot software, a Github Copilot competitor, launched a host of new features and raised lots of money for its advancement further into AI.

Hugging Face releases its own version of ChatGPT - Hugging face is an AI startup, and they recently released Hugging Chat, an open-sourced alternative to ChatGPT. It isn’t as powerful as ChatGPT, but is free and has ambitions to be much bigger.

Let me know what you think. And, as always, please share!

Share

With the introduction of ChatGPT, Dall-E, and a host of companies and products, we’re entering a new age of AI, automation, and technology.

If you haven’t been keeping up, don’t feel bad. Every day there is news about new advances, new products, new companies, and new applications for AI. It is more than anyone can keep up with.

So what to do?

In this newsletter, we’ll be exploring news from the week, along with what is coming down the pike for AI and other technologies. I’ll be exploring how I use new products in my work and how you can start to use them in yours.

We’ll talk with CEOs and founders of companies, explore the implications of new technology (good and bad), and what the next horizon looks like.

If you’re ready, let’s jump in. And I’d love it if you shared this newsletter (and upcoming podcast) with others.

Subscribe now

Share

Subscribe now