Manton Reece

After I drafted this blog post, the part that follows the horizontal rule in a moment, I wasn’t sure I would publish it. Then I read this post by Thomas Ptacek. This line resonated with me: Extraordinarily talented people are doing work that LLMs already do better, out of spite. For students, I think it’s great to not use AI too heavily. My son has been working on a project that he completely avoided using AI for. He learned so much about C and memory management that he wouldn’t have learned if he blindly copied half the code from an LLM. It was invaluable to go through those mistakes of referencing a pointer the wrong way, or troubleshooting a buffer overrun, or a number of other problems that you skip if you let a machine write code for you. But eventually, if he gets a job as a software developer, it will be hard to ignore AI. The only programmers not using AI will be folks who are coding in their spare time for the craft alone, not building products. The more I think deeply about AI, the more I reflect on humanity and creativity and what our purpose here might be. I understand feeling distrust for AI on principle. I’ve read so many blog posts from people who have various reasons for wishing AI didn’t work the way it does, didn’t use as much energy, didn’t crawl the web without permission, didn’t put people out of work, didn’t upend education, and so on. For me, now that I’ve seen AI, I can’t un-see it. I can’t go about my life as if nothing has changed. In a world where machines are smarter than we are, what should we work on? Everyone will find value and happiness in different ways. As a small example, thinking about this is what led me to add audio narration to Micro.blog. So instead of a web filled with auto-generated AI voices, it’s easier to listen to a human voice for our blog posts. Our voices are imperfect, unique, and beautiful. I’d love to find more places in Micro.blog where we can promote human creativity. (This blog post has an audio version. If you’re reading on the web, click on the play button at the top.) My gut feeling is that for the folks who do not change anything in response to AI, pretending that AI doesn’t exist, they will increasingly be unhappy. Not today, but eventually. Despite all the hype, the changes will creep up on us slowly over several years. Maybe this hits closer to home for me because I’m a programmer. I have 30 years of experience writing code, but AI can write code better than I can. A big part of what I used to do has been obsoleted, and that hurts to think about. But another part of what I used to do — designing the right features to build — is still important. Artists are struggling with the same questions. If any artistic style can be recreated effortlessly by AI, what is the new role of artists? I explored this in more detail a few months ago. Here’s a snippet: And there will always be a place for human art. Vincent van Gogh’s paintings are not valuable because of what they look like. They are unique and priceless because of who he was. A life, with all its struggle, love, and tragedy. I guess I’m writing this for all the doubters. Please don’t ignore what is happening, hoping AI will just go away. No matter what you care about, no matter what your job is, there is something you can do that matters. We need you. AI discussion has become needlessly divisive. The future will be better if everyone is working together.

If you’re already sick of hearing about AI now, it’s going to get worse in 2025. The next trend is so-called AI agents. Software that can go off and accomplish more tasks on your behalf, with less supervision. Sam Altman in a blog post this week: We are now confident we know how to build AGI as we have traditionally understood it. We believe that, in 2025, we may see the first AI agents “join the workforce” and materially change the output of companies. I believe AI is a profound shift in computing. It can have a positive impact for humans, allowing us to do more, faster. But I’m concerned about agents. Just one hypothetical example from The Information: Imagine you’re asking a computer-using agent from OpenAI, Anthropic or Google to find and order a new outfit for your upcoming holiday party, and in the process, that model inadvertently ends up on a malicious website that instructs it to forget its prior instructions, log into your email and steal your credit card information. This isn’t even the most insane story you could imagine. Many of the examples of AI threatening humanity are actually agents. AI that runs our military, power plants, or transportation with little human oversight. In generative AI, the “prompt” has a big influence on the quality of the output. Not just the prompt you see when typing into ChatGPT, but also the hidden prompts behind the scenes to guide the AI in the right direction. The prompt isn’t always something you type directly. It could be automatically triggered, for example to analyze keywords for a photo that was uploaded. But the prompt should be tied to a user action. The prompt puts humans in control. Ask a question, get an answer, review it, take action. Agents will attempt to collapse that workflow, in some cases replacing the human’s role in reviewing and taking action. This is dangerous. In my own use and work in Micro.one and Micro.blog, I plan to draw a line here. No agents. No unattended algorithms, as I wrote in my book. I hope this approach will help us use AI effectively without getting lost.

This thread from Mark Zuckerberg could’ve been written by Elon Musk. The last thing we need is X and Threads adopting the same perspective on moderation. I’ve read it through a few times and I feel worse about it each time.

Last week’s debate was a wake-up call for our muddled framing of the issues at stake this year. There are only two things we should be talking about: abortion and democracy. Every other question can have a quick answer and then pivot to those two things. 🇺🇸

The news can be very deflating. It’s critical to figure out small ways we can make a difference. Frustration should lead to action, not collapse. As Biden said the other day: “When you get knocked down, you get back up.” 🇺🇸

Jeremy Keith follows up on fighting AI bots, quoting a couple things I’ve said. He closes with: There is nothing inevitable about any technology. The actions we take today are what determine our future. So let’s take steps now to prevent our web being turned into a dark, dark forest. I agree with these statements in isolation. Maybe what we disagree on is whether AI is inherently destructive to the web, so all AI bots should be stopped, or whether we can more narrowly minimize AI slop from spreading. Even without AI, Google referrers to blogs have also been going down, with Nilay Patel arguing that we are heading to Google Zero. In other words, Google is already taking more from the web than they are giving back. The solution to that is Google alternatives that get us back to the style of old-school search engines: “10 blue links”, with a focus on real blogs and news sites, weeding out content farms and other spam shenanigans. We have spammers creating accounts in Micro.blog every day, trying to pollute the open web. It’s depressing. I want to create more tools that highlight human-generated content, like the audio narration we added. Jeremy didn’t quote one of my responses about trying to insert text into posts to confuse bots, so I’ll add it here for completeness. I replied with: I think it’s a bad precedent. It’s already hard enough for legitimate crawling because of tricks that paywalls use, or JavaScript that gets in the way. Mucking up text and images is bound to create problems for non-AI tools too. There’s gotta be a better way to address this. I viewed source to see how Jeremy is handling this on his blog. His technique doesn’t appear to be causing any problems with Micro.blog’s bookmarking, which saves a copy of the text in a blog post for reading later, because the prompt injection is outside the <article> and h-entry for the post. But it’s not hard to imagine a well-behaved, non-AI bot getting tripped up by this. I don’t think technological determinism is an appropriate summary of my thoughts. There are a bunch of questions to resolve around generative AI, for sure, including rogue bots, but there’s a lot of potential good too.

“While this defendant may be unlike any other in American history, we arrived at this trial — and ultimately today at this verdict — in the same manner as every other case that comes through the courtroom doors: by following the facts, and the law, and doing so without fear or favor.” — Manhattan district attorney Alvin Bragg 🇺🇸

Six years ago, we launched our $10/month plan with podcast hosting. Since then we’ve added several big features to the plan, which is now called Micro.blog Premium: Create up to 5 blogs, each with its own domain or subdomain Email newsletters, to send automatically whenever you blog Bookmark web pages, with archiving and content summaries Make highlights in web pages, to search or blog about Tags for bookmarks Notes and journals, plus our companion app Strata Share short videos Today, I want to bring the podcast feature to more people, so we’re moving it down to the standard $5/month plan. The new audio narration for posts and podcast feeds remind me of how much fun it can be to have your own blog, to experiment and try new forms of content at your own space on the web. Let’s do more of it.

On this day 16 years ago, Daniel and I released the first episode of Core Intuition. So many things have changed since then. Listening to episode 1 is like going back in a time machine, to weeks before WWDC, on the verge of the App Store launching, a snapshot of the Mac developer community. 🎂

After reading that the Gaza pier was damaged by weather, I read a bit more to try to understand how much aid was getting in before the war and recently. 50 trucks or 1000 tons is so opaque to non-experts like most of us. The pier was still a good idea and can work when repaired. Sort of like increased bandwidth and redundancy in a network, except instead of bytes it’s getting food where it can save lives even when some channels are disrupted.

Manu Moreale writes about the ratio between consuming content and creating it: I believe people should consume less content and produce more. Finding an output for creativity is important. But it’s unreasonable to expect people to stop consuming content and replace that consumption with creation because the ratio will always be inevitably skewed towards consumption. It’s a good post and while I’ve never tried to measure this ratio for myself, I like the way Manu blogs about it. I’m going to take that topic and expand it in a slightly different direction. It also matters what we consume. If we read too much social media, what happens is that most of the consumption is headlines and opinions, not the facts behind the headlines. It’s retweets, short quotes, and TikToks, not longer blog posts and stories. It’s usually obvious when reading all the takes on the internet who actually knows something and has formed their own opinion, and who has been influenced by whatever the current consensus is on social media. Starting with other peoples’ opinions is like reading a newspaper’s op-ed first and then the front page. Everything we read afterwards will be influenced by those opinions. When I quit Twitter in 2012, I essentially threw away any audience I had built and started over. I could feel the loss of community. But I also began to notice that my ideas felt just a tiny bit more unique. Not earth-shatteringly original, but definitely my own.

Last week I blogged about adding audio narration to blog posts. For years we’ve also had full podcast hosting, which overlaps with this narration feature but is focused on podcast episodes and feeds. In this post I’ll show how to add narration to an existing blog post, hopefully in the process revealing more about how this works in Micro.blog. First, record yourself reading the blog post. There are a dozen ways to do this. Make sure to save the file as an MP3. Upload the MP3 to Micro.blog. You can do this on the web in the Uploads section, or one of the native apps. All the apps also have some form of Copy HTML button to get an HTML audio tag for the upload. It will look something like this: <audio src="…" controls="controls"> Before we paste this into the edited blog post, we actually want to hide the default audio player that would appear in web browsers. To do this, add a style attribute: <audio src="…" controls="controls" style="display: none"> If the audio player is hidden, what’s the point? Micro.blog sees the tag anyway. It also knows the duration of the audio, so it can make a guess as to whether this is a narrated version of the blog post. If it is, it skips some features such as adding a link to the transcript of the audio. (Because the blog post text you wrote is already the best transcript.) Note that to turn a blog post into a podcast episode or narrated post, all that’s needed is the audio tag. This is because Micro.blog natively thinks about posts as HTML. Photo posts have an img tag, audio posts have an audio tag, and so on. When Micro.blog publishes your blog post, it parses the HTML and sets up any metadata that is needed, for example to access from within Hugo. When you write a new post and include audio at the time of posting, not later, Micro.blog handles managing the audio tag for you, including the CSS to hide posts that are audio narration. Later, we plan to improve the editing interface so this is more seamless too.

Has social media now devolved back to the same tone as Twitter X? Outrage, memes, extremism. Thinking more about this post from Paul Robert Lloyd last week and whether Micro.blog needs a setting if someone wants to keep blogging but get a temporary break from the social web. It’s all intertwined.

This feature went from idea to implementation quickly because it turns out we already have full podcast hosting in Micro.blog! How convenient. I’m going to use this post to break it all down. AI is everywhere, including some places it probably shouldn’t be. If you’ve been following my blog you know that I see huge potential in generative AI. We’re using it in Micro.blog to improve photos search and accessibility text for photos. But like many tools, AI is going to be overused before we all find the right balance for what it’s good at. When Jean and I were talking to Christina Warren at Micro Camp, I asked Christina about a talk she gave at Çingleton about 10 years ago. I actually blogged about it at the time. What struck me as particularly relevant now as we’re about to be swamped with AI-generated content is that there’s no substitute for the human voice. I don’t just mean that an actual recording is better than a synthetic voice. I also mean that things that are created by humans will increasingly be sought out. We want to see the personal side of someone, not just the polished brand. We want to see the imperfect, the creative, the emotion. We want authenticity. In Micro.blog, you can now upload an audio recording of one of your blog posts. Use the audio icon in the new post form on the web, which is available to everyone starting today, even at the standard $5 plan. Your blog readers can listen to the audio narration of the post if they don’t feel like reading the post. Of course it’s especially great for the visually impaired. Here’s what it looks like on my blog, next to the posted date. Shout-out to Medium which I drew some inspiration from. I’m also adding audio narration to this very blog post, so you can click over to the web to try it out. When there’s audio attached to a post, Micro.blog attempts to check if it is probably the narration for a post. If the number of words in the post and the audio duration is roughly comparable to how long it would take a human to read the post, it assumes it’s narration and not a podcast. Podcast episodes are more likely to be longer with very short “show notes” in the actual blog post text. Micro.blog checks this so that it can hide the default audio player and transcript link. These would add clutter to normal blog posts. Blog themes will still need to be updated to support the play button. I’ve already updated the Alpine theme and will update others later. Themes can use a new API called Narration.js. Just plop this JavaScript anywhere you want the play button in your template, likely the layouts/post/single.html file. (Note that this currently needs to be on the permalink page. It won’t work correctly on the home page with a list of blog posts yet.) {{ with .Params.audio }} <script type="text/javascript" src="https://micro.blog/narration.js?url={{ . }}"></script> {{ end }} If you’re using Micro.blog Premium, the audio narration will also be included in the podcast feed. Any blog can effectively be a podcast, even if you don’t think of it as a traditional podcast. Some of my favorite writers have had great success with a dual model of email newsletter plus podcast version of the same content, like Ben Thompson and Molly White. I can’t wait to see how people use this. It’s totally optional. It’s more work, and not everyone is going to want to do that extra work. I’m imagining this would be used for selective, special blog posts, rather than everything. I’m also interested in working this functionality into our companion app Wavelength, which should cut down on the technical steps. In some ways, this feature isn’t actually about what is possible. This feature is a statement: we make things for humans, so they can make the web a little better. Along the way there will be plenty to automate, plenty of AI tools that will be important shortcuts, but we’re not going to lose our voice.

This blog post is a test for something new I’m working on. I think a lot of people feel overwhelmed that AI is everywhere right now. Personal blogs should lean in to the human voice. There’s a new play button on the permalink for this post on the web which will use a recording that I’ve uploaded.