Martin Hearn

👻 Kiro 👻 has had a whole host of great new features recently. There are the really big ones like Skills support and the constant evolution with new models like Opus and Sonnet 4.6, but one of my favourites is ACP... The Agent Client Protocol lets you take advantage of one of the best coding assistants 💪 around and use the underlying agent in a variety of different clients. I've been exploring how you can take advantage of Kiro for use cases that aren't all about writing code 💻 and how you and your teams can all reap the benefits from its huge capability. I've got my latest post in the comments on some of my favourite new ways to use Kiro with ACP. Let me know what you think. 🚀 #kiro #genai #kirocli

27

2 Comments

Your dev team says “consent mode’s implemented.” Your conversions say otherwise. Abbie Poots (founder of Perazam Consulting Ltd) breaks down why it still might be set up wrong and how to check. 🎧 Episode 8 out now. https://lnkd.in/ePA8FiB6 #WebMarketerPodcast #ConversionTracking #ConsentMode #B2BMarketing

5

“Don’t optimise early” is *right* for software But “don’t optimise ever”? Only if you want bigger bills and more expensive power I’m looking at 5 - controversial - ways software engineering could do better this week. Number 1 is “optimisation”. The maxim most people know is “Don’t optimise early!” And that’s right. Otherwise you’ll waste time and money almost all of the time. But “don’t optimise EVER”? A big mistake. These days, many people feel that it’s cheaper, easier, quicker (pick any you want) to just throw more cloud resources at a problem. And some application spaces, like LLM construction. often rely directly on pure brute force. (Honourable mention to all teams developing more cost-conscious approaches though!) Because - and this isn’t just about AI and LLMs - hugely increasing cloud usage and forecast usage is… ⚠️ driving up electricity prices [plus factors too, of course] ⚠️ creating water scarcity ⚠️ costing more money! What’s your take - is optimisation a thing of the past, or should it be growing in importance?

17

14 Comments

Anything, a software development company founded by Marcus Lowe, has acquired the Anything.com domain name in a massive upgrade. The domain name had long been used as a brand by a domain investment portfolio company. https://lnkd.in/ed7j5ntZ

20

1 Comment

Following on from my previous post about Glastonbury’s use of data from app ‘favourites’ to predict and manage crowds, similar approaches are in use at Wimbledon. Their long-standing tech partner, IBM, have really grasped the value of this customer behavioural data. Like Glastonbury, the Wimbledon app allows fans to favourite players, and this can be used to help forecast court attendance, supporting decisions on staffing, crowd flow and safety. Digital engagement data within the app also provides valuable signals. Spikes in app and web traffic often align with key match moments. These patterns can help identify when audiences are most engaged and when to deliver content or support services. Social media sentiment is analysed in real time. Comments about queues, weather, or app issues are reviewed and acted on quickly. This provides a live feedback loop. These are just a few examples that show how data can support decisions to improve the customer experience. Often, the data already exists - it’s about knowing it's there, and putting it to use to deliver value. What other examples of this have you seen?

10

2 Comments

At Gateway X, we prefer to be close to the cutting edge of AI—but not 𝑜𝑛 it. We’ve all got kids at home, businesses to run, and who knows what else (the water line to the fridge just broke and sprayed water all over). We couldn't keep up even if we wanted to. There are lots of side benefits to this posture... We get: 💡 𝘁𝗵𝗲 𝗹𝗮𝘁𝗲𝘀𝘁 𝘁𝗼𝗼𝗹𝘀 — Stuff from 2 months ago is still 𝑏𝑟𝑎𝑛𝑑 𝑛𝑒𝑤. Most businesses are way further behind. ⚙️ 𝘁𝗵𝗲 𝗼𝗻𝗲𝘀 𝘁𝗵𝗮𝘁 𝗮𝗰𝘁𝘂𝗮𝗹𝗹𝘆 𝘄𝗼𝗿𝗸 — We let the market filter out what’s “neat” but not useful. Dodged a few hype traps this way. 🤖 𝘀𝘆𝘀𝘁𝗲𝗺𝘀 𝘁𝗵𝗮𝘁 𝗱𝗼𝗻’𝘁 𝗯𝗿𝗲𝗮𝗸 𝗲𝘃𝗲𝗿𝘆 𝘄𝗲𝗲𝗸 — New = inexperience. Let it simmer a bit, and you learn how to 𝑡𝑎𝑙𝑘 to these things. That’s when tools turn into real workflows. — 🔥 𝗛𝗢𝗧 𝗧𝗜𝗣: When we build automations, we reference 𝑑𝑎𝑡𝑒𝑑 versions of models (e.g., `gpt-4.1-2025-04-14` instead of just `gpt-4.1`) So model updates don’t suddenly ruin our outputs 💥 — Here’s what’s working for us today 👇 𝗖𝗵𝗮𝘁𝗚𝗣𝗧 🧠 `o4-mini` – default for fast chats w/ some extra thought. Slower than 4o, but sharper. 🐢 `o3` – verrry slow. But deep. When I want to think, not talk. 𝗔𝗣𝗜 / 𝗢𝗽𝗲𝗻𝗪𝗲𝗯𝗨𝗜 ⚡ `4.1` – fast + reliable. My default for new tools. 🧘 `o3` again for careful, slower work. 𝗚𝗲𝗺𝗶𝗻𝗶 2.5 𝗣𝗿𝗼 📄 1M token context window (!!) 👂 Great for dumping in 30 transcripts and pulling out real signal. 🚀 Surprisingly good out of the box. — Some quirks: ✍️ System prompts help `4.1` a 𝑙𝑜𝑡. You can get very specific. 🤯 But they hurt `o3` / `o4-mini`. Too much structure throws them off. 🛠️ That’s why `4.1` runs our automations—it's a better listener. Why do we default to OpenAI? Sure, we use other platforms too, but it's nice to have a "home". Other platforms are just more things to learn, more to keep up with, and they are all leap frogging each other every month or two anyway. — Just what’s working for us right now. What’s working for you?

16

US Tariffs appear to be back for UK and European brands. Previously there was a way to mitigate these if a brand has a US entity by moving stock wholesale between entities. MAUVAIS launched on Patchworks in just 14 days to mitigate the last set of tariffs. In fast changing global economics, a flexible, agile tech stack is a must. Read more here: https://lnkd.in/eKZvecrE

47

10 Comments

**Phygital. Remember that word?** It popped up after everyone got bored of webrooming and showrooming (or even boomerooming – a term I coined back in 2017 to prove it was never a one-way journey). It peaked around 2023 when every brand promised "seamless digital-physical experiences". Then it quietly died, because most of it was marketing theatre. Screens in stores. QR codes that broke. "Scan to buy online" gimmicks. But the idea was right. The execution wasn’t. We’re now at the point where phygital can actually work, not as an "Augmented reality" sideshow but as part of the payments infrastructure itself. Reality? - Tapping your payment card in a store and it instantly recognises you, links to your digital account, and syncs stock, loyalty, fulfilment, and identity in one move. Buy in person, fulfil online. The Infinite Aisle made real. That’s phygital commerce, built on real rails: payments, identity, and data. And it’s possible today. For companies like Amazon, Apple, or any major marketplace, the missing piece has always been the in-person link. Aevi’s world is built around exactly that, connecting physical touchpoints to digital platforms through one consistent set of pipes. So yes, we’re bringing Phygital back. Only this time, it’s actual infrastructure....

6

1 Comment

Founders - swapping the agency model for an in-house team? 👉 Here's the 3 things you can’t afford to wing I’ve worked with some terrific agencies that feel like true partners But plenty are just build shops You ask, they code, job done Which works - until it doesn’t Your app is live First ten customers are happy enough Then you hire your first engineer and everything unravels 🚫 No product docs 🚫 All the “why” buried in Zoom recordings 🚫 AWS login? Who knows 🔍 Your new hire spends a month playing archaeologist ⛏️ Digging through code 🧑‍🦯 Chasing Slack threads 🤞 Praying nothing breaks Heading into this handover? Nail these basics first ✅ Write down the story PRD, user journeys, key decisions. Get them somewhere visible Dump voice notes into ChatGPT or Notion AI and let it draft the first pass ✅ Walk through your cloud Get shown where it lives, how to spin it up, how to shut it down Even if you don’t grok every toggle, you’ll know where the bodies are buried ✅ Ask for the safety net Unit tests, end-to-end tests, release checklist If “our tests live in the developers’ heads” is the answer, stop the handover until that’s fixed Clean handover now Fewer 3 am fire drills later Going through this transition now? 👉 Let me know how it's going in the comments

19

9 Comments

I have written some new RFC (request for comments) for EventCatalog. After speaking and collecting feedback from many companies and users, there are themes that come up. So I spent some time understanding them and writing some updates on how they would work with EventCatalog - Webhooks (e.g setup notifications on schema changes) - Channel Chaining (Support channel > channel > channel) - Let domains send/receive messages (making services optional) - Support Data Products (warehouses, ML, dashboards) - Add Containers (external systems, databases, complete architecture) Which would be useful for you? What am I missing? RFC Links below, if you have time to vote or review any, that would be awesome ❤️

16

4 Comments

To niche or not to niche? 🤔 Our konversable tech is now in place with clients in over 12 different sectors. This goes against the common thinking in tech which is to make sure you “niche-down” on your product offering. But truthfully, if we had done this, I don’t think we would be here now. If you have a product which is specifically designed to solve a pain point in a particular sector, then of course it makes sense to focus on that. But that doesn’t mean to say all tech should follow the same principle. We have managed to find footholds with our solution across many different sectors, for many different use cases. Frankly that’s one of the reasons I love our tech so much, in that it can be applied in so many different ways to solve clients problems. So next time you are told that you must “niche-down”, don’t take the advice at face value, really consider if it’s worth doing, or if, as in our case, broadening your offering across sectors will allow you to scale much quicker. (Photo of me with a cuppa has nothing to do with this post but LI algo will prefer this to an AI generated gif 😂) #konversable #saas

32

3 Comments

Apologies for another marketing whinge but please, please don't ask me to find time in my diary for "a 20 minute demo of your product". a) It's never 20 minutes b) if it were relevant I would absolutely reach out and ask you for a demo on the strength of a brief explanatory email and c) if I agreed to even half of these that would be my whole week taken care of. Worse though, when I decline, asking me who in my or organisation you should contact is such a no no. Firstly asking 'who deals with audience platforms' or 'who deals with streaming' is a nonsense request in a large organisation. Secondly, finding the right person in many cases would result in an intolerable amount of research and thirdly I'm not giving out colleagues emails without their consent.

159

31 Comments