Yasith Fernando

I know Silicon Valley startups don't want to hear this..... But the combination of someone in the trades with deep domain expertise and Claude Code will run circles around your generic software. I talked to Cory LaChance this morning, a mechanical engineer in industrial piping construction in Houston. He normally works with chemical plants and refineries, but now he also works with the terminal He reached out in a DM a few days ago and I was so fired up by his story, I asked him if we could record the conversation and share it. He built a full application that industrial contractors are using every day. It reads piping isometric drawings and automatically extracts every weld count, every material spec, every commodity code. Work that took 10 minutes per drawing now takes 60 seconds. It can do 100 drawings in five minutes, saving days of time. His co-workers are all mind blown, and when he talks to them, it's like they are speaking different languages. His fabrication shop uses it daily, and he built the entire thing in 8 weeks. During those 8 weeks he also had to learn everything about Claude Code, the terminal, VS Code, everything. My favorite quote from him was when he said, "I literally did this with zero outside help other than the AI. My favorite tools are screenshots, step by step instructions and asking Claude to explain things like I'm five." Every trades worker with deep expertise and a willingness to sit down with Claude Code for a few weekends is now a potential software founder. I can't wait to meet more people like Cory.

Introducing 𝑨𝒕𝒕𝒆𝒏𝒕𝒊𝒐𝒏 𝑹𝒆𝒔𝒊𝒅𝒖𝒂𝒍𝒔: Rethinking depth-wise aggregation. Residual connections have long relied on fixed, uniform accumulation. Inspired by the duality of time and depth, we introduce Attention Residuals, replacing standard depth-wise recurrence with learned, input-dependent attention over preceding layers. 🔹 Enables networks to selectively retrieve past representations, naturally mitigating dilution and hidden-state growth. 🔹 Introduces Block AttnRes, partitioning layers into compressed blocks to make cross-layer attention practical at scale. 🔹 Serves as an efficient drop-in replacement, demonstrating a 1.25x compute advantage with negligible (<2%) inference latency overhead. 🔹 Validated on the Kimi Linear architecture (48B total, 3B activated parameters), delivering consistent downstream performance gains. 🔗Full report: github.com/MoonshotAI/Att…

These days whenever I use any SaaS, I ask myself: can claudecode replace this or make this substantially better? And the answer is almost always yes. SaaS might not be dead but I reckon the products of tomorrow will be quite different

wow it's so cool that they added our favorite feature from the Claude Code CLI to the desktop app

Things that were not on my bingo card 1. Anthropic marked as US supply chain risk by the government (same as Huawei (!!)) 2. Claude going viral across non-tech people as a result, #1 on the app store 3. Mass ChatGPT cancellations thanks to OpenAI becoming the gov't AI supplier

Amazing hackery!

@simonw Both

YouTube question: I've been making a few videos recently and I'm torn between the honest no-cheating live coding approach and the here's-something-I-prepared-earlier approach I'm aiming for 10-30 minutes per video Which format do people find more useful?

@Domon Did this actually happen?!

Boss: Why didn't you do the thing? Me → Claude: Why didn’t you? Claude → Subagent: Why didn’t you? Subagent: ...

Excuse for new laptop 🤣

This should be good

I suspect the companies who have decided to reduce or fix headcount with AI are doing it out of fear. Perhaps because they don't know if they can win. The AI native companies are hiring like crazy.

low-fi wireframes are the best (these are from codex-cli)

Yo I heard if u press Up, Up, Down, Down, Left, Right, Left, Right, B, A in Sam Fransisco there's an infinite money glitch

@sarmadsangi Veo3 can also do Sri Lankan accents really really well. Which I found super interesting given how little training data they have compared to more popular languages

been experimenting with Veo3 and TIL that … Veo3 can extract accent from an image apparently when generating a video. This is very interesting but weird if you were born without an accent (in another country).

@sarmadsangi Assuming you also give it text with the image, what if you describe the person’s background as American Indian etc? May be it’ll then use a more American accent. Or even something more specific like south bay born and raised Indian American etc

This is incredible

It's cool to have first party support for both Claude and GPT5 in the same tool. And it potentially unlocks new features for Cursor CLI that single provider tools can't support.

One of the biggest downsides of Claudecode is that there is no first party support for using non Anthropic models. While Cursor CLI/Agent has the unique advantage of supporting many models. This is one competitive advantage Cursor has.

Love all the competition ClaudeCode is getting. CLI is a good form factor coding agents.