One of the most impactful uses of LLM technology lies in content rewriting, which naturally capitalizes on these systems’ robust capabilities for generating and refining text. This application is a logical fit, helping users enhance their content while engaging with a service.

Similar to summarization but incorporating an element of judgment, features like Microsoft Team CoPilot’s call transcript summaries distill extensive discussions into essential bullet points, spotlighting pivotal moments or insights.

The ability to ‘understand’ nuanced language through summarization extends naturally into advanced search functionalities. ServiceNow does this by enabling customer service agents to search tickets for recommended solutions and to dispel jargon used by different agents.

Rather than merely focusing on content creation or manipulation, emerging applications of these systems provide new perspectives and predict outcomes based on accumulated human experiences. The actual value of these applications lies not merely in enhancing efficiency but in augmenting effectiveness, enabling users to make more informed decisions.

what if every police patrol car had a bank of cameras that scan not just every number plate but every face within a hundred yards against a national database of outstanding warrants? What if the cameras in the subway do that? All the connected cameras in the city? China is already trying to do this, and we seem to be pretty sure we don’t like that, but why? One could argue that there’s no difference in principle, only in scale, but a change in scale can itself be a change in principle.

Generative AI is now creating a lot of new examples of scale itself as a difference in principle. You could look the emergent abuse of AI image generators, shrug, and talk about Photoshop: there have been fake nudes on the web for as long as there’s been a web. But when high-school boys can load photos of 50 or 500 classmates into an ML model and generate thousands of such images (let’s not even think about video) on a home PC (or their phone), that does seem like an important change. Faking people’s voices has been possible for a long time, but it’s new and different that any idiot can do it themselves. People have always cheated at homework and exams, but the internet made it easy and now ChatGPT makes it (almost) free. Again, something that has always been theoretically possible on a small scale becomes practically possible on a massive scale, and that changes what it means.

This might be a genuinely new and bad thing that we don’t like at all; or, it may be new and we decide we don’t care; we may decide that it’s just a new (worse?) expression of an old thing we don’t worry about; and, it may be that this was indeed being done before, even at scale, but somehow doing it like this makes it different, or just makes us more aware that it’s being done at all. Cambridge Analytica was a hoax, but it catalysed awareness of issues that were real

The ability of generative AI tools to analyse huge datasets can also be used to help spark creative inspiration. This can help us if we’re struggling for time or battling writer’s block. For example, if a social media manager is looking for ideas on how to engage alumni on Instagram, they could ask ChatGPT for suggestions based on recent popular content. They could then pick the best ideas from ChatGPT’s response and adapt them. We may use these tools in a similar way to how we ask a colleague for an idea on how to approach a creative task.

We may use these tools in a similar way to how we use search engines for researching topics and will always carefully fact-check before publication.

we will not publish any press releases, articles, social media posts, blog posts, internal emails or other written content that is 100% produced by generative AI. We will always apply brand guidelines, fact-check responses, and re-write in our own words.

We may use these tools to make minor changes to a photo to make it more usable without changing the subject matter or original essence. For example, if a website manager needs a photo in a landscape ratio but only has one in a portrait ratio, they could use Photoshop’s inbuilt AI tools to extend the background of the photo to create an image with the correct dimensions for the website.

inside look at how Perplexity builds product—which to me feels like what the future of product development will look like for many companies:AI-first: They’ve been asking AI questions about every step of the company-building process, including “How do I launch a product?” Employees are encouraged to ask AI before bothering colleagues.Organized like slime mold: They optimize for minimizing coordination costs by parallelizing as much of each project as possible.Small teams: Their typical team is two to three people. Their AI-generated (highly rated) podcast was built and is run by just one person.Few managers: They hire self-driven ICs and actively avoid hiring people who are strongest at guiding other people’s work.A prediction for the future: Johnny said, “If I had to guess, technical PMs or engineers with product taste will become the most valuable people at a company over time.”

Typical projects we work on only have one or two people on it. The hardest projects have three or four people, max. For example, our podcast is built by one person end to end. He’s a brand designer, but he does audio engineering and he’s doing all kinds of research to figure out how to build the most interactive and interesting podcast. I don’t think a PM has stepped into that process at any point.

We leverage product management most when there’s a really difficult decision that branches into many directions, and for more involved projects.

The hardest, and most important, part of the PM’s job is having taste around use cases. With AI, there are way too many possible use cases that you could work on. So the PM has to step in and make a branching qualitative decision based on the data, user research, and so on.

a big problem with AI is how you prioritize between more productivity-based use cases versus the engaging chatbot-type use cases.

we look foremost for flexibility and initiative. The ability to build constructively in a limited-resource environment (potentially having to wear several hats) is the most important to us.

We look for strong ICs with clear quantitative impacts on users rather than within their company. If I see the terms “Agile expert” or “scrum master” in the resume, it’s probably not going to be a great fit.

My goal is to structure teams around minimizing “coordination headwind,” as described by Alex Komoroske in this deck on seeing organizations as slime mold. The rough idea is that coordination costs (caused by uncertainty and disagreements) increase with scale, and adding managers doesn’t improve things. People’s incentives become misaligned. People tend to lie to their manager, who lies to their manager. And if you want to talk to someone in another part of the org, you have to go up two levels and down two levels, asking everyone along the way.

Instead, what you want to do is keep the overall goals aligned, and parallelize projects that point toward this goal by sharing reusable guides and processes.

Perplexity has existed for less than two years, and things are changing so quickly in AI that it’s hard to commit beyond that. We create quarterly plans. Within quarters, we try to keep plans stable within a product roadmap. The roadmap has a few large projects that everyone is aware of, along with small tasks that we shift around as priorities change.

Each week we have a kickoff meeting where everyone sets high-level expectations for their week. We have a culture of setting 75% weekly goals: everyone identifies their top priority for the week and tries to hit 75% of that by the end of the week. Just a few bullet points to make sure priorities are clear during the week.

All objectives are measurable, either in terms of quantifiable thresholds or Boolean “was X completed or not.” Our objectives are very aggressive, and often at the end of the quarter we only end up completing 70% in one direction or another. The remaining 30% helps identify gaps in prioritization and staffing.

At the beginning of each project, there is a quick kickoff for alignment, and afterward, iteration occurs in an asynchronous fashion, without constraints or review processes. When individuals feel ready for feedback on designs, implementation, or final product, they share it in Slack, and other members of the team give honest and constructive feedback. Iteration happens organically as needed, and the product doesn’t get launched until it gains internal traction via dogfooding.

all teams share common top-level metrics while A/B testing within their layer of the stack. Because the product can shift so quickly, we want to avoid political issues where anyone’s identity is bound to any given component of the product.

We’ve found that when teams don’t have a PM, team members take on the PM responsibilities, like adjusting scope, making user-facing decisions, and trusting their own taste.

What’s your primary tool for task management, and bug tracking?Linear. For AI products, the line between tasks, bugs, and projects becomes blurred, but we’ve found many concepts in Linear, like Leads, Triage, Sizing, etc., to be extremely important. A favorite feature of mine is auto-archiving—if a task hasn’t been mentioned in a while, chances are it’s not actually important.The primary tool we use to store sources of truth like roadmaps and milestone planning is Notion. We use Notion during development for design docs and RFCs, and afterward for documentation, postmortems, and historical records. Putting thoughts on paper (documenting chain-of-thought) leads to much clearer decision-making, and makes it easier to align async and avoid meetings.Unwrap.ai is a tool we’ve also recently introduced to consolidate, document, and quantify qualitative feedback. Because of the nature of AI, many issues are not always deterministic enough to classify as bugs. Unwrap groups individual pieces of feedback into more concrete themes and areas of improvement.

High-level objectives and directions come top-down, but a large amount of new ideas are floated bottom-up. We believe strongly that engineering and design should have ownership over ideas and details, especially for an AI product where the constraints are not known until ideas are turned into code and mock-ups.

Big challenges today revolve around scaling from our current size to the next level, both on the hiring side and in execution and planning. We don’t want to lose our core identity of working in a very flat and collaborative environment. Even small decisions, like how to organize Slack and Linear, can be tough to scale. Trying to stay transparent and scale the number of channels and projects without causing notifications to explode is something we’re currently trying to figure out.

Some images look like they were made under a robotic sugar high. Lots of warm colors, but they make everything look like candy… they’re so overly sweet that they give you visual diabetes..

Average AI images drag down everything around them. An AI hero image is a comedian opening the show with a knock-knock joke. Good images enrich your article, bad images steal its soul.

Forget trying to decide what your life’s destiny is. That’s too grand. Instead, just figure out what you should do in the next 2 years.

Visuals can stir up feelings or paint a scene in an instant. However, they may not always nail down the details or explain things as clearly as words can. Words can be very precise and give you all the information you need. Yet, sometimes they miss that instant impact or emotional punch.

For each visual you add to your presentation, you should ask yourself “What does it really say?” And then check: Does it enhance the meaning of my message, or is it purely decorative? Does it belong at this point in my presentation? Would it be better for another slide? Is there a better image that says what I want to say?

Computers don’t feel, and that means: they don’t understand what they do, they grow images like cancer grows cells: They just replicate something into the blue. This becomes apparent in the often outright creepiness of AI images.

AI is really good at making scary images. Even if the prompt lacks all hints of horror kitsch, you need to get ready to see or feel something disturbing when you look at AI images. It’s like a spell. Part of the scariness comes from the cancer-like pattern that reproduces the same ornament without considering its meaning and consequence.

Placing pictures next to each other will invite comparisons. We also compare images that follow each other. Make sure that you do not inadvertently compare apples and oranges.

When placing multiple images in a grid or on one slide after the other, ensure they don’t clash in terms of colors, style, or resolution. Otherwise, people will focus more on the contrast between the images rather than their content.

Repeating what everyone can see is bad practice. To make pictures and text work, they need to have something to say about each other.

Don’t write next to the image what people already see. A caption is not an ALT text.

The most powerful combination of text and image happens when the text says about the image what you can’t see at first sight, and when the image renders what is hard to imagine.

Do not be boring or overly explanatory. The visual should attract their attention to your words and vice-versa.

If a visual lacks meaning, it becomes a decorative placeholder. It can dilute your message, distract from what you want to say, and even express disrespect to your audience.

XML tags are used to wrap around content, like this: <tag>content</tag>. Opening and closing XML tags should share exactly the same name. The tag name can be anything you like, as long as it's wrapped in angle brackets, although we recommend naming your tags something contextually relevant to the content it's wrapped around.

To get the most out of XML tags, keep these tips in mind: Use descriptive tag names that reflect the content they contain (e.g., <instructions>, <example>, <input>). Be consistent with your tag names throughout your prompts. Always include both the opening (<tag>) and closing (</tag>) tags, including when you reference them, such as Using the document in <doc></doc> tags, answer this question. You can and should nest XML tags, although more than five layers of nesting may decrease performance depending on the complexity of the use case.

if you had showed VisiCalc to a lawyer or a graphic designer, their response might well have been ‘that’s amazing, and maybe my book-keeper should see this, but I don’t do that’. Lawyers needed a word processor, and graphic designers needed (say) Postscript, Pagemaker and Photoshop, and that took longer.

I’ve been thinking about this problem a lot in the last 18 months, as I’ve experimented with ChatGPT, Gemini, Claude and all the other chatbots that have sprouted up: ‘this is amazing, but I don’t have that use-case’.

A spreadsheet can’t do word processing or graphic design, and a PC can do all of those but someone needs to write those applications for you first, one use-case at a time.

no matter how good the tech is, you have to think of the use-case. You have to see it. You have to notice something you spend a lot of time doing and realise that it could be automated with a tool like this.

Some of this is about imagination, and familiarity. It reminds me a little of the early days of Google, when we were so used to hand-crafting our solutions to problems that it took time to realise that you could ‘just Google that’.

This is also, perhaps, matching a classic pattern for the adoption of new technology: you start by making it fit the things you already do, where it’s easy and obvious to see that this is a use-case, if you have one, and then later, over time, you change the way you work to fit the new tool.

The concept of product-market fit is that normally you have to iterate your idea of the product and your idea of the use-case and customer towards each other - and then you need sales.

Meanwhile, spreadsheets were both a use-case for a PC and a general-purpose substrate in their own right, just as email or SQL might be, and yet all of those have been unbundled. The typical big company today uses hundreds of different SaaS apps, all them, so to speak, unbundling something out of Excel, Oracle or Outlook. All of them, at their core, are an idea for a problem and an idea for a workflow to solve that problem, that is easier to grasp and deploy than saying ‘you could do that in Excel!’ Rather, you instantiate the problem and the solution in software - ‘wrap it’, indeed - and sell that to a CIO. You sell them a problem.

there’s a ‘Cambrian Explosion’ of startups using OpenAI or Anthropic APIs to build single-purpose dedicated apps that aim at one problem and wrap it in hand-built UI, tooling and enterprise sales, much as a previous generation did with SQL.

Back in 1982, my father had one (1) electric drill, but since then tool companies have turned that into a whole constellation of battery-powered electric hole-makers. One upon a time every startup had SQL inside, but that wasn’t the product, and now every startup will have LLMs inside.

people are still creating companies based on realising that X or Y is a problem, realising that it can be turned into pattern recognition, and then going out and selling that problem.

A GUI tells the users what they can do, but it also tells the computer everything we already know about the problem, and with a general-purpose, open-ended prompt, the user has to think of all of that themselves, every single time, or hope it’s already in the training data. So, can the GUI itself be generative? Or do we need another whole generation of Dan Bricklins to see the problem, and then turn it into apps, thousands of them, one at a time, each of them with some LLM somewhere under the hood?

The change would be that these new use-cases would be things that are still automated one-at-a-time, but that could not have been automated before, or that would have needed far more software (and capital) to automate. That would make LLMs the new SQL, not the new HAL9000.

Living in an immigrant, multilingual family will open your eyes to all the ways humans can misunderstand each other. My story isn’t unique, but I grew up unable to communicate in my family’s “default language.” I was forbidden from speaking Korean as a child. My parents were fluent in spoken and written English, but their accents often left them feeling unwelcome in America. They didn’t want that for me, and so I grew up with perfect, unaccented English. I could understand Korean and, as a small child, could speak some. But eventually, I lost that ability.

I became the family Chewbacca. Family would speak to me in Korean, I’d reply back in English — and vice versa. Later, I started learning Japanese because that’s what public school offered and my grandparents were fluent. Eventually, my family became adept at speaking a pidgin of English, Korean, and Japanese.

This arrangement was less than ideal but workable. That is until both of my parents were diagnosed with incurable, degenerative neurological diseases. My father had Parkinson’s disease and Alzheimer’s disease. My mom had bulbar amyotrophic lateral sclerosis (ALS) and frontotemporal dementia (FTD). Their English, a language they studied for decades, evaporated.

It made everything twice as complicated. I shared caretaking duties with non-English speaking relatives. Doctor visits — both here and in Korea — had to be bilingual, which often meant appointments were longer, more stressful, expensive, and full of misunderstandings. Oftentimes, I’d want to connect with my stepmom or aunt, both to coordinate care and vent about things only we could understand. None of us could go beyond “I’m sad,” “I come Monday, you go Tuesday,” or “I’m sorry.” We struggled alone, together.

You need much less to “survive” in another language. That’s where Google Translate excels. It’s handy when you’re traveling and need basic help, like directions or ordering food. But life is lived in moments more complicated than simple transactions with strangers. When I decided to pull off my mom’s oxygen mask — the only machine keeping her alive — I used my crappy pidgin to tell my family it was time to say goodbye. I could’ve never pulled out Google Translate for that. We all grieved once my mom passed, peacefully, in her living room. My limited Korean just meant I couldn’t partake in much of the communal comfort. Would I have really tapped a pin in such a heavy moment to understand what my aunt was wailing when I knew the why?

For high-context languages like Japanese and Korean, you also have to be able to translate what isn’t said — like tone and relationships between speakers — to really understand what’s being conveyed. If a Korean person asks you your age, they’re not being rude. It literally determines how they should speak to you. In Japanese, the word daijoubu can mean “That’s okay,” “Are you okay?” “I’m fine,” “Yes,” “No, thank you,” “Everything’s going to be okay,” and “Don’t worry” depending on how it’s said.

The “hard tech” in AI are the LLMs available for rent from OpenAI, Anthropic, Cohere, and others, or available as open source with Llama, Bloom, Mistral and others. The hard-tech is a level playing field; startups do not have an advantage over incumbents.

There can be differentiation in prompt engineering, problem break-down, use of vector databases, and more. However, this isn’t something where startups have an edge, such as being willing to take more risks or be more creative. At best, it is neutral; certainly not an advantage.

This doesn’t mean it’s impossible for a startup to succeed; surely many will. It means that you need a strategy that creates differentiation and distribution, even more quickly and dramatically than is normally required

Whether you’re training existing models, developing models from scratch, or simply testing theories, high-quality data is crucial. Incumbents have the data because they have the customers. They can immediately leverage customers’ data to train models and tune algorithms, so long as they maintain secrecy and privacy.

Intercom’s AI strategy is built on the foundation of hundreds of millions of customer interactions. This gives them an advantage over a newcomer developing a chatbot from scratch. Similarly, Google has an advantage in AI video because they own the entire YouTube library. GitHub has an advantage with Copilot because they trained their AI on their vast code repository (including changes, with human-written explanations of the changes).

While there will always be individuals preferring the startup environment, the allure of working on AI at an incumbent is equally strong for many, especially pure computer and data scientsts who, more than anything else, want to work on interesting AI projects. They get to work in the code, with a large budget, with all the data, with above-market compensation, and a built-in large customer base that will enjoy the fruits of their labor, all without having to do sales, marketing, tech support, accounting, raising money, or anything else that isn’t the pure joy of writing interesting code. This is heaven for many.

A chatbot is in the chatbot market, and an SEO tool is in the SEO market. Adding AI to those tools is obviously a good idea; indeed companies who fail to add AI will likely become irrelevant in the long run. Thus we see that “AI” is a new tool for developing within existing markets, not itself a new market (except for actual hard-tech AI companies).

AI is in the solution-space, not the problem-space, as we say in product management. The customer problem you’re solving is still the same as ever. The problem a chatbot is solving is the same as ever: Talk to customers 24/7 in any language. AI enables completely new solutions that none of us were imagining a few years ago; that’s what’s so exciting and truly transformative. However, the customer problems remain the same, even though the solutions are different

Companies will pay more for chatbots where the AI is excellent, more support contacts are deferred from reaching a human, more languages are supported, and more kinds of questions can be answered, so existing chatbot customers might pay more, which grows the market. Furthermore, some companies who previously (rightly) saw chatbots as a terrible customer experience, will change their mind with sufficiently good AI, and will enter the chatbot market, which again grows that market.

the right way to analyze this is not to say “the AI market is big and growing” but rather: “Here is how AI will transform this existing market.” And then: “Here’s how we fit into that growth.”

Thinking in dialogue is easier and more entertaining than struggling with feelings, letters, grammar and style all by ourselves. Using AI as a writing dialogue partner, ChatGPT can become a catalyst for clarifying what we want to say. Even if it is wrong.6 Sometimes we need to hear what’s wrong to understand what’s right.

Seeing in clear text what is wrong or, at least, what we don’t mean can help us set our minds straight about what we really mean. If you get stuck, you can also simply let it ask you questions. If you don’t know how to improve, you can tell it to be evil in its critique of your writing

Just compare usage with AI to how we dealt with similar issues before AI. Discussing our writing with others is a general practice and regarded as universally helpful; honest writers honor and credit their discussion partners We already use spell checkers and grammar tools It’s common practice to use human editors for substantial or minor copy editing of our public writing Clearly, using dictionaries and thesauri to find the right expression is not a crime

Using AI in the editor replaces thinking. Using AI in dialogue increases thinking. Now, how can connect the editor and the chat window without making a mess? Is there a way to keep human and artificial text apart?

Instead of turning light into pictures, it turns ideas into pictures.

This new kind of camera replicates what your imagination does. It receives words and then synthesizes a picture from its experience seeing millions of other pictures. The output doesn’t have a name yet, but I’ll call it a synthograph (meaning synthetic drawing).

Photography can capture moments that happened, but synthography is not bound by the limitations of reality. Synthography can capture moments that did not happen and moments that could never happen.

Taking a great syntho is about stimulating the imagination of the camera. Synthography doesn’t require you to be anywhere or anywhen in particular.

Photography is an important medium of expression because it is so accessible and instantaneous. Synthography will even further reduce barriers to entry, and give everyone the power to convert ideas into pictures.