Activity
Mon
Wed
Fri
Sun
Sep
Oct
Nov
Dec
Jan
Feb
Mar
Apr
May
Jun
Jul
Aug
What is this?
Less
More

Owned by Simon

AI Dev Academy

142 members • Free

Master AI-assisted development with Claude Code. Learn from real projects, network with other developers, and get hired by The AI Dev Agency.

Showcasing AI for Business

48 members • Free

Mobility & Injury Prevention

239.2k members • Free

AI Made EASY!

59 members • Free

Skoolers

161.5k members • Free

AI Automation Society

437.5k members • Free

AI Automation Agency Hub

332k members • Free

AI with Mike

17 members • Free

Synthesizer: Free Skool Growth

45.9k members • Free

39 contributions to AI Dev Academy
Fable 5
Anyone using it? I have maximised it to the full, on phase 2 & 3 of a client build! Now out of credit until Monday afternoon. With a small window of opportunity, I’m planning over the weekend to use the full availability before it goes to a paid plan. Are you using it? How are you finding it? How are you planning to take advantage of it while it’s available?
0
0
AI Agent Memory Is Not Chat History
Hey everyone 👋 I just published a short article about something I keep thinking about while building Agent Systems: "AI Agent Memory is Not Chat History". The main idea: giving agents more context is not the same as giving them trustworthy memory. Real memory needs scope, provenance, freshness, permissions, and evidence. I would love to hear how people here are handling memory in their own agents. Thanks! Article here: https://dev.to/glendel/ai-agent-memory-is-not-chat-history-4jjb
0 likes • Jun 29
Hey Glendel, Finally got to this. The core idea is genuinely great. Your points on authority hierarchy, multi-agent propagation and workflow state belongs to the runtime, not the model are sharp and correct. The diagnosis is rock solid. But you stop just as it gets interesting. There are some inconsistency in the sections, and layers overlap and don't have clear defendable boundaries. This is a repacking of MemGPT, the whole 24-25 context-engineering wave, but it's not named. You could shorten it considerably without losing the main idea. You could followup with some insights on treatment, one concrete working example, a sketch of an enforcement mechanism. Looking forward to the next one!
Ik zoek een Nederlander.
Ik ben van plan om met mijn ontwikkelteam de Nederlandse markt te betreden. We staan ​​echter voor een aantal uitdagingen, te beginnen met de taalbarrière. Ik ben momenteel op zoek naar iemand die in Nederland woont en mij wil ondersteunen. Programmeervaardigheden zijn een pluspunt, maar niet vereist. Beheersing van meerdere talen is bijzonder wenselijk. Dit is geen kortlopend project; ik zoek iemand die een langetermijnvisie deelt. Het gaat hier niet om gratis werk. We kunnen de details via DM bespreken. Ik stel een snelle reactie op prijs.
0 likes • Jun 29
Good Luck! Wish you every success!
The quiet wasn't empty
Last week I told you why I'd gone quiet for five months. A few of you asked what I was actually doing. Here's the real version, and it's messier than any highlight reel. For about three months, almost nothing happened that you could see. Under the surface, it was bugs, then more bugs, then refactoring, then pushing my engineering further than I ever had. Crisis management at 3 am. Long stretches where the only win was that something finally stopped breaking. Meanwhile, the accelerator I'm in went wild for Claude Code. Everyone shipping, everyone talking, and in the space of a few days, I felt weeks behind. That feeling is brutal. Then I actually looked at what people were building, and it hit me. I wasn't behind at all. I was a long way in front. I'd just been too head-down to notice. Not all of it felt like progress. Someone was testing a new product of mine, and weeks later, I found out they hadn't really used it and had no feedback on what was wrong with it. That stung. So I rebuilt how I validate things: proper auditing, tracking, testing, so I'm never that exposed again. I also won a bid against someone whose cost of living is a fraction of mine. The only way to win that is to be so much better that price stops being the question, which meant pushing the engineering to the edge again to stay ahead. Then the pattern every solo founder knows. Nothing for three months, and then everything at once. Three new coding contracts landed (one fell through), two new joint ventures opened up, and I had to put the code down entirely to deal with contracts. My main salon client threw up its own challenges too, including the oldest question in delivery: when is it actually done? I changed tactics, it worked, and I moved on to the next fire. Here's the thing I want you to take from this. Those quiet months weren't lost time. They were the foundation on which the loud ones were built. If nothing is visible to you right now, that might not be a problem at all. It might be the groundwork.
2
0
The Sound Of Silence
Wow! My last post in here was the 29th of Jan. It was about Clawbot. It all seems like a distant dream now! That's almost five months of silence, and here's the uncomfortable reason why. I've been heads-down building. Client work, contract negotiations, crisis management (thanks for starting a war, Mr President, right next to where my main client is), deadlines, and possible Joint Ventures, the actual things I get paid for. By the time the day ends, writing a post about how well it's going was the last thing on my mind. For a while that made me feel like I was falling behind. Everyone else seems to post every day. Then I looked at who "everyone" actually is and I realised something interesting. Two kinds of people post constantly. The first kind has no work. Plenty of time, no clients, maybe a half-baked, half interesting idea, but mostly they post about the dream: the mindset, the morning routine, the AI news, or that model that they used on this todo list project, or that launch that's always two weeks away. The second kind has loads of work, but they've handed most of it off. They built a team, bought their time back, and now they post from the top of the hill. Both have time. That's the trick. Your feed is full of these two groups because they're the only ones with a free hour to talk. The people in the middle go quiet. The ones with real clients and no team yet, doing the work with their own two hands (and the helpful - sometimes - assistance of their AI model of choice). They're not posting because they're busy doing the thing the other two are talking about. That's where I've been for five months. Middle of the climb. No spare hour, plenty of proof. More on this in another post! So if you feel behind because you're not posting daily, read that again. Your silence might be the most honest signal you've got. It means you're in the work while other people are still talking about it. Faking it at the empty end gets you nothing. The aim is to earn your way to the far end, where you've handed enough off to win your time back and you've got something real to point at.
1-10 of 39
Simon Filtness
3
21 points to level up
@psy-phi-4882
Tech leader turned startup founder. Helping AI Agency owners and SME owners build custom AI Automations and Solutions

Active 2d ago
Joined Aug 18, 2025
United Kingdom