CodeWithStu

Principal Engineer building in public (@onehausapp). Previously CTO @ Illuvium, Principal @ Checkout dot com, Lead Architect @ ClearBank

[CodeWithStu @CodeWithStu]

Hey hey 👋 Lots of new people recently, so let's get to know each other

I'm Stu. I'm a principal engineer who has spent years at places like Checkout and ClearBank. Whilst I'm building my future, I share interesting things and what I learn on the way

What are you working on?

[CodeWithStu @CodeWithStu]

Just waiting for Apple to do this distribution thing, so building out MCP connection capabilities to my app so you can use it from Hermes or Open Claw or ChatGPT or Gemini or Claude. What’s app coming soon too!

[CodeWithStu @CodeWithStu]

@theo Also i think people put too many skills in place and then wonder why 50k of tokens are used for a single prompt

[CodeWithStu @CodeWithStu]

@theo I think a lot of people also don’t know what can help manage the inference costs as well. Using things like single type per file (less tokens on read), tools like LSP instead of file searches and other tools like RTK to reduce the token count of things like git

[CodeWithStu @CodeWithStu]

First thought: wow

Second thought: isn’t there a ton of caching that OpenAi et al could do and reduce the burden? not sure if in place already or not

[Matthew Prince 🌥 @eastdakota]

Welp, that happened faster than I predicted. Thought it would be end of 2027, then early 2027, but agentic traffic growing so fast that bots have now passed human traffic online for the first time in the Internet's history. https://t.co/2zX5bHdhsa

[CodeWithStu @CodeWithStu]

@SkyNews Aren’t we basically broke as a country and we funding this and hs2? 😂

[Google Gemma @googlegemma]

Meet Gemma 4 12B!

A unified, encoder-free multimodal model designed to bring high-performance intelligence directly to your laptop, and released under an Apache 2.0 license.

Bridging the gap between edge efficiency and advanced reasoning. Here is what’s new with Gemma 4 12B: 👇

[CodeWithStu @CodeWithStu]

I did the professsional voice clone and i did an hour of recording i think it was. But the result is that it cloned my voice perfectly. More data = better result. Main thing i had to do was to prompt their MCP to fix certain words to have their phonetic spelling rather than actual, eg: sequel not SQL

[Hayden Barnes @unixterminal]

BLOG: WSL 2 is getting faster Windows file system access https://t.co/xb6xvke6lA

[CodeWithStu @CodeWithStu]

@MythThrazz Not bad, not bad :)

[CodeWithStu @CodeWithStu]

Some really interesting data here

[Tim Soulo 🇺🇦 @timsoulo]

In the last 6 months at @Ahrefs, we analyzed over 1 billion data points across 14 studies. Here's what we learned about AI search optimization:

1) "Best X" blog listicles are the single most prominent content format cited by AI chatbots. They make up 43.8% of all page types cited by ChatGPT specifically.

2) 67% of ChatGPT's top 1,000 citations come from sources marketers can't influence: Wikipedia (29.7%), homepages (23.8%), app stores (6.6%). Only 32.3% are influenceable content like educational pages, reviews, news, and blog posts.

3) 28.3% of ChatGPT's most-cited pages have zero Google organic visibility. These pages get cited repeatedly by ChatGPT despite not ranking in Google at all. A completely separate discovery layer.

4) ChatGPT only cites about 50% of the URLs it retrieves. It fetches dozens of pages per query but uses half as background context without attribution. This means that being retrieved and being cited are very different things.

5) Adding schema markup had zero meaningful impact on AI citations. AI Overviews actually dipped −4.6%, while AI Mode (+2.4%) and ChatGPT (+2.2%) showed changes indistinguishable from zero.

6) YouTube mentions have the highest correlation (0.737) with AI brand visibility out of all the factors we studied (including all the conventional SEO metrics like backlinks, page count, DR, etc). This held true for both Google-owned and OpenAI products.

7) AI Overviews reduce clicks to the #1 result by 58%. That’s up from 34.5% just 10 months earlier. The trend is accelerating.

8) 99.9% of AI Overviews appear on informational intent queries. Transactional, navigational, and local searches are almost entirely AIO-free. Shopping triggers AIOs just 3.2% of the time.

9) For a given search query, Google’s AI Mode and AI Overviews reach the same conclusions 86% of the time — but cite almost entirely different sources (only 13.7% citation overlap).

10) AI Overviews change every 2.15 days on average, with 70% of content differing between consecutive observations. But semantic similarity stays at 0.95. The words, sources, and entities constantly shuffle, but the actual meaning barely moves.

[CodeWithStu @CodeWithStu]

@sandislonjsak Here’s an example. If i added “Add a task” to the front of that, the LLM would find the task tool and implement. Or i can click the task button and the traditional UI appears

[CodeWithStu @CodeWithStu]

I think it's because most companies are saying "use our chat it's awesome", not "connect our thing to your chat" eg: via MCP, skills, CLI.

The choice should be the users of which harness they use and which tools they want to allow that access to. For the majority that will be ChatGPT or Claude. It won't be OpenClaw/Hermes/Eliza/NetClaw or any other ones. Users (non-techies) want 1-click deploy and trust not a random install script.

Remember, it's something crazy like 80% of people have never opened an AI tool like ChatGPT. Still very early

[Sandi Slonjšak @sandislonjsak]

Seems to me like we are close to chatbot fatigue.

Users do not want to chat with every product.

Sometimes they just want the annoying task to disappear without having a conversation about it.

[CodeWithStu @CodeWithStu]

@sandislonjsak 100% - that's why with my app, I've got both things. You can use the chat box to do stuff OR you can use the classic UIs

[CodeWithStu @CodeWithStu]

@thdxr How about taking them to a coldplay concert?

[CodeWithStu @CodeWithStu]

@SimonHoiberg Looks good!

I missed that you added X back. It was the only reason i didn’t use it when i was automating heavily. Totally understand why though the pricing was insane

[CodeWithStu @CodeWithStu]

@marclou Awesome. Can’t wait to try this out soon

[CodeWithStu @CodeWithStu]

https://t.co/D3b9bOm5ez

[CodeWithStu @CodeWithStu]

Whilst building @onehausapp - i came across an interesting idea. I could allow embedding my app as a what’s app bot which could be a huge uplift in productivity for some families. Only reason i haven’t to date is because of the per message charge would eventually eat away any margin i had, i think.

What would fellow builders do here?

[CodeWithStu @CodeWithStu]

Working on my mascot for @onehausapp marketing. Using my IRL dog as inspiration. Generated using a local model and I'm impressed.

[CodeWithStu @CodeWithStu]

@bas_fijneman It's just a matter of time and finding the right people. Just watched the content in the library in @shipordie_ - some really good stuff in there. I can definitely see a dedicated IG account doing well hell with the visuals