Toni Hopponen

Have you ever decided not to add a product image to a page just because it's so hard to find and update them later? Well, I have 😢 So, I built a tool that helps me browse, search, and replace images in help docs, blog posts, etc. with a few clicks 👇

@aipush_app @BeCitable totally. From a content writer's pov, the challenge sometimes is that AI would love question + answer style, but converting humans from content in an optimal way might have other requirements

@tonihopponen @BeCitable The missing piece most people skip: ChatGPT doesn't crawl like Google. It needs static HTML, clear structure, one answer per page, and explicit AI crawler permissions. Content alone won't get you cited — the technical foundation decides if AI even reads you.

How do I get my service mentioned in ChatGPT? It's a question every founder and marketer has been asking themselves this year. I interviewed Cole from @BeCitable to find out how to find content ideas and write stuff that get mentions 👇

Watch and listen to the full episode in your favourite app: YouTube: youtube.com/watch?v=Wflcq9… Spotify: open.spotify.com/episode/4KbVKF… Apple Podcasts: podcasts.apple.com/gb/podcast/b2b…

@rileycx I haven’t tried Webflow specifically, but for my site, I have an MCP that finds the right page, the right part of the text (or visual component), and updates it for me. A waste of time in most content update scenarios when the editor is AI-powered

@tonihopponen like via Webflow MCP?

hot take: it doesn't matter if your site is built in webflow, framer, wix, or even squarespace (which i say through gritted teeth). they all output completely usable code that works fine in search, can be accessible, etc. i've had conversations with founders who built on squarespace and had great organic traffic and their site looked great. the code is not the problem. for context — i run a web design studio. we've launched a ton of webflow sites and a ton of framer sites. we're actively building client projects in both platforms right now. this isn't me saying these tools are bad because i can't use them. i know them well and they're great at what they do, especially if you are a technical user and know how to build websites. but the only real difference between any of these tools is developer experience. how easy is it to build on, how easy is it for a founder or marketing team to maintain. that's the whole conversation. and if that's what we're actually optimizing for — ease of building and ease of maintenance... then there's no reason to use anything other than a codebase that can be queried and updated through an ai chat. no drag-and-drop editor. no platform lock-in. just your code, hosted wherever you want, updated by talking to it. and you own everything.

Opus 4.6. is following our past behaviour. Problem? Let's add some more tools to our stack. Just 10 or 100x, of course

@NapierHolland crashes on my phone 😅

They actually hired a copywriter! Solid work. Observe how easily you can skim this homepage thanks to clear headlines that summarise each section. They also poke fun at the status quo. 👍

@darrenmarble brutal

Describe being a founder in one word.

Take this weekend off.

@Remotion is quickly becoming one of my favourite marketing tools. In ~10 mins, I get simple videos done for micro marketing things like promoting a new feature 👇 Not replacing the video editing tools like Veedio for me, but giving me a completely new marketing skill.

@madiou @_baretto Same here. SEO tools always feel expensive vs the value I get

@_baretto Some of these I have never heard of ngl but I'm guessing Ahrefs

Here's every single tool we pay for: Cloud Canon Woodpecker Drag App Crisp Resolve247 GSuite Loom Fathom Flaticon Ahrefs Chartmogul CodeRabbit AugmentCode Guess which costs the most...

"I wish we had a page for this" Now, the solution is just one Slack message away 👇

Vibecoding a video with Claude and Remotion is pretty fun. The first try 👇

@helloitsolly There's a bit of work with all the tools needed + learning where everything's located in the codebase, but many tasks are combining or expanding existing assets in the codebase. For the past 6 months or so, I've been only building and shipping the code to my co-founder

Currently I use v0 for new feature wireframes And the engineer uses cursor for fixing bugs reported to linear But I’m wondering if I should start product changes in cursor and then hand them over to the engineer Will this create more problems than it solves? Really new to using ai like this

@NapierHolland @otter_ai Granola

What's your favourite AI call transcription service? @otter_ai is totally unreliable — it shits the bed on massive, important conversations.

Who wants me to audit their B2B SaaS/AI homepage messaging? Will add the first 5 replies to my list

@forgebitz One day at a time. For 5-10 years 😅

the best business advice i can give you: focus stop fcking around with random shit and trying to hit 100k mrr in a week, it's going to take time and dedication don't get distracted

@geirfreysson Yesterday, I was thinking aloud about how some of the logic in our code isn’t easy to explain. If this, then that. Feels like the current level performs so well if you can give a task like "Look at this, follow the same, add this". Maybe should optimise for that says non-dev 🙋

Talking to devs who’ve never used Claude Code (or similar) feels like a blue pill/red pill moment.

@omgitsdef More is always better. But only with the quality button turned on 😅

There's a fallacy going on in the startup/VC world that more content = better. I see Heads of Marketing who say they're shipping 5 blogs a week and want to double that. Then I look at the blogs they've published, and they're absolute garbage. No one reading them gets any value from them. Those 10 shitty blog posts would have been better spent writing 1 blog post driven by actual insights that AI can't replicate. And with AI and the investment into AI is only recirculating this lie that you can ship 100 pieces of high-quality longform content at a time. When the data shows that AI favors high-quality, insight-driven pieces. Not the garbage you're cranking out with AI. It's time for the startup world to look in the mirror and ask itself: Why are we doing this? Is there a point? The answer's probably no