What a shitty SaaS billing experience @PageFlows. Last year you sent me a reminder 1 week out from the renewal - perfect 👌 Silly me expected the same again and didn't get it 🙄 Actually, I haven't had a single email from you since that reminder a year ago... since then you've arbitrarily decided I don't need billing reminders? Complete silence from you over the last year until you charge my card. Why not send reminders to all people on all plans. Why not keep is consistent from year to year?

Travelling at the moment. If we cross paths and you want to meet up. Give me a holler. KL 🇲🇾 (now ~ 15th) Da Nang 🇻🇳 (15th ~ 22nd) HCMC 🇻🇳 (22nd ~ 29th) SIN 🇸🇬 (29th & 30th)

The real key here is leveraging AI to your advantage to 10x your output, not trying to fight it so you can continue to do what’s comfortable, which is doing things the way you’ve always done them - manually. Anecdotal feedback on coding assistants for example suggests that they take a 10x engineer and make them a 100x engineer. They don’t replace them, they give them superpowers. I suspect in just about every industry this will hold true, even for designers. I don’t think anyone can dispute that a skilled designer can do amazing work that AI just isn’t capable of, but at what cost? A skilled designer might spend a whole day producing something that AI can output at 80% in minutes. That’s 80% of the value at maybe 0.5% of the cost, and I get it now and not next week. The harsh reality is that is probably an awesome outcome in a lot of cases - job done. Just look at the impact Canva has had? The designer that flourishes is the one that takes that 80% and gets it to 120% in a couple of hours. Now that’s compelling! I have been impacted by big tech shifts in the corporate world in the past. Honestly, people need to get over themselves. This has been happening in other industries for decades and those that don’t adapt end up out the door. Those that do become indispensable. I’ve seen many once high-flyers from great companies end up out the door, and then endlessly bouncing backwards and downwards to companies that are steps behind until the music finally stops. Don’t do this!

We’re definitely entering an age where big walls are starting to cast big shadows onto the garden. Rather than those walled gardens being the sanctuaries they once were, the sun is going to set quickly over 2025 for those platforms that don’t take a more open and collaborative approach. It’s important to be an enabler, rather than get in the way. The rate of change is like nothing we’ve seen before, and it’s only accelerating. Quite frankly it’s overwhelming for even the best resourced and best funded companies. IMO the best approach is to unlock the gate so the community can build upon your foundations in a way you could never. It’s natural to want to protect your sanctuary, but those that don’t adapt and continue to guard the gate may very well find themselves standing in a cold, dark and empty withered garden before too long. I’m constantly reminded of this video these days, which demonstrates that just because something was clear as day yesterday, it doesn’t mean it’ll still be true tomorrow.

It was only a matter of time before AI coding assistants started adding visual editing controls for fine tuning. I suspect they’ll continue to push quickly towards no-code style visual editors focusing initially on UI, and at some point we’ll probably see a logic tree.

I just got off a call with a long time client who wants to take some time to recalibrate their strategy moving forward (B2B SaaS intelligence tool) given all of the activity in the AI space over the last month. I completely agree with them! I don't think the UI layer will disappear anytime soon, but it can be massively simplified with multi-modal agents and generative UI in a lot of cases. Why build a dashboard for someone and put the onus on them to log in and interpret raw data when you can ask them what they are focused on / concerned about that the moment and send them some custom analysis at regular intervals or when certain metrics hit thresholds that they can then click into and have a conversation about.

Has anyone gone in to optimise a @bubble app and it gets stuck on the loading screen? How did you solve it? 1. I've asked support about this previously and after much confusion and back and forth they gave some generic response to get me to go away. 2. My computer / internet bandwidth has loads of headroom so it's not on my end. I suspect Bubble is struggling to compile the list and eventually times out getting stuck on the loading screen. 3. We've been working on this project for a while after another team had built it out over a couple of years. The app bundle is huge and I suspect the optimisation has never been run as there is loads of deleted fields, options and so on. We're getting super weird bugs where something is correctly configured and works for a while, but then randomly throws an error and the only resolution is to delete that workflow and rebuild it exactly as it was before. Rinse and repeat. It's getting a little tedious not to mention a shitty UX and this is probably the last thing left to try. Although it clearly needs to be done anyway. Any insights would be amazing 🙏

😂😂😂

Deploying on the weekend - testing in production, naturally!

AI tools are awesome. I use several on the daily. But this couldn’t be more true… Just like an actual person, no model knows everything about everything, so it’s not possible for them to fire 100% of the time. I often ask for simple things like code for an API call, even from major well documented services like Stripe. They commonly invent features or endpoints that don’t exist and the calls don’t work. Even with feedback loops they can’t always dig themself out of the hole. Don’t get me wrong, they’re awesome, but if you have no idea at all about what you’re asking, you’ll end up in a terrible mess most of the time.

I'm working remotely from 🇹🇭Bangkok for this next week. Anyone I should have a coffee with? Then I'll be in 🇰🇷 Seoul from the 10th & 🇯🇵Tokyo from the 17th.

Me 😅: @stripe is it possible to increase some account limits on this account, as the current limits exclude almost all transactions? Stripe 💳: tell us what you need and we'll look into it. 📅 2 weeks pass Stripe 💳: you need to build up more transaction history before we can do anything for you Me 😕: It's not even worth setting up the integration & business processes for a tiny % of transactions that are so infrequent. How will this even justify an increase down the line? Ok, let's try this - the business is 13 years old doing $100k+ MRR, just not through stripe. What do you need to be comfortable and I'll get it? Stripe 💳: you need to build up more transaction history before we can do anything for you. --- Why do companies even bother having a support function if all they do is make me wait weeks to get a canned response that doesn't even try to apply any sort of common sense or at least acknowledgement of what I'm asking? I would have been satisfied even if the response was "our policy is pretty strict, it's a bit of a stalemate in this case, maybe you'll need to look elsewhere" Might as well just put in support AI and give me an instant answer vs. waiting two weeks for someone to copy/paste and hit send.

The infinite recursion protection feature by @bubble is an awesome idea until you realise there is no consideration for time. If you have a maximum depth of 10 (the default setting) and you have a workflow that schedules itself to re-run a month later, it would fail rescheduling itself on month 10 as the limit has been reached. Absolutely crazy 🤯 Wouldn't a better implementation be to choose the maximum depth & a time limit. Or if that's too difficult, a static 24 hours? I'm almost certain the issue that most people have are infinite loops that kick in and burn through 100,000 or even millions of WU in hours, and not the sleepy old workflow that ticks along each day or even month. This means that you end up needing to set the limit far higher than you otherwise may need to, negating the benefits of the feature in the first place.

In @bubble , how would you go about generating a list of date ranges from a list of dates? The use case is starting with a date range (say July 1 - 30) but there could be multiple blackout dates in that period. So you might end up with a handful of ranges like this: 1-10 12-15 20-24 27-30 How would you go about that?

Was a great chat with @twesolowski