Davor Pihač

FusionCache is now featured on the official Redis docs in the Libraries and tools section 🥳 redis.io/docs/latest/in…

What a strange take. Oxfmt is drop in Prettier replacement and it takes couple of minutes to setup. You can try it and decide if it works for you. Gains are massive and risk is minimal. I'll take faster any time.

@dailydotdev Switched in a couple of minutes and haven’t had any regrets. Big performance gains resulted in faster pipelines and shorter LLM loops

oxfmt is 30x faster than prettier. most devs will stick with prettier because we're tired of switching tools every 3 months for nothing

I just released FusionCache v2.6.0 🎉 In this version: 🏷️ configurable RemoveByTag() behavior Ⓜ️ RemoveByTag("*") support in HC adapter 🔒 better distributed locker + Eager Refresh ⚡ perf boosts 🐞 minor bug fixes github.com/ZiggyCreatures…

This is probably the lowest I've been since a few years. Can't find freelance work, was super close of getting a good job in a startup in Switzerland, I failed B2B sales, and I don't know how I'll pay rent this month. Worst of all: my second baby is coming next month. I did tons of outreach that I should have started earlier. But it'll likely won't bring any result for at least a month. So yeah, if you're looking for a senior developer with 10+ years of experience (18 in total, 13 as a pro), tons of skills, Laravel + ecosystem expert, LLMs expert, design expertise, business expertise, etc. Feel free to reach out.

What is crazy is actually owning it, being human and respectful. Content over form.

@AlexFinn Except that monthly payments and electricity bill will be much higher than $200 per month. Hardware cost of that machine would be at least $10k and would not make sense for most individuals. Companies would make sure to utilize machine 24/7

Apple will win the AI war in June 2026 without investing a single penny into AI The war won without a sword They will release the M5 Ultra with 1TB of memory It will be able to run super intelligence more powerful then Opus locally on your desk for free It will fundamentally change the way the entire world operates. EVERYONE with a personalized super intelligent AI employee working on their desk 24/7 $200 a month AI plans will feel antiquated While you've been crapping on Apple for not entering an AI race against 100 other companies shredding trillions of dollars, they ran their own race Personalized local super intelligence will be stage 1 of the singularity. I sincerely hope you're preparing for this

I am the Senior Product Manager for Notepad at Microsoft. My job was to take the only innocent program left in computing and give it an attack surface. I succeeded. CVE-2026-20841. CVSS 8.8. Remote code execution. In Notepad. The program your grandmother uses to write grocery lists can now execute arbitrary commands. I got promoted. Notepad was invented in 1985. For forty years it did one thing. It opened text. It did not render text. It did not format text. It did not connect to the internet. It did not have opinions about your writing. It was the last honest program on your computer. My team fixed that. In 2024 we killed WordPad. "Nobody uses it," we said. "It's bloat," we said. Then in May 2025 we added Markdown rendering to Notepad. In September we added AI-powered writing, rewriting, and summarization. We turned Notepad into WordPad. Then we turned it into Word. Then we turned it into an attack vector. The roadmap was always the same. We just changed the name on the headstone. The vulnerability works like this: you open a Markdown file in Notepad. You click a link. Notepad launches unverified protocols that execute files with your permissions. A text editor. Running code. From a link. In a program that existed specifically to NOT run code. We shipped this as a default feature. The "off" switch is in settings. Nobody checks settings. We know this. We designed it that way. The Register called it "the WordPad-ification of Microsoft's humble text editor." Critics said we were betraying Notepad's core ethos. They were correct. Notepad's ethos was simplicity. Simplicity does not generate engagement metrics. Simplicity does not require a Copilot+ PC. Simplicity does not produce CVEs. We gave Notepad AI, network access, Markdown rendering, and tables. We gave it everything except a threat model. Forty years of doing nothing wrong. We couldn't allow that to continue. This is the Microsoft way.

@jodydonetti Congratulations 🎉

FusionCache has been approved by the European OpenSource project 🎉

Engineers report up to 50% more productivity when using AI. Code generation, debugging, testing and documentation are now all being generated with AI. Productivity of individual contributors has skyrocketed, but what nobody talks about is overall team productivity, quality and lead time. AI doesn’t eliminate communication overhead, hierarchies or slow decision and approval processes. Large, inefficient teams don’t automatically become effective just because individuals generate code more quickly. The important question is not whether engineers are more productive. It’s whether clients and end users experience faster delivery, better quality, and shorter lead times.

@SevenviewSteve I’m using Svelte with ShadCN Svelte and Bits UI. It has taken me significant amount of time to create proper instructions. Now it mostly works fine. Similar happens with C# output, which is not as good as I would have liked (still quite usable).

I had Claude Code build the same UI in 5 different stacks. React, Hotwire, Inertia+Vue, Inertia+React, and vanilla HTML/JS. Same spec, same features. The quality gap between frameworks was massive. Wrote up what happened and what I think it means for picking a stack in 2026. x.com/SevenviewSteve…

@HeroDividend Time spent writing and reading this email cost more than 3 minutes in total 😅

What would you do if you got this email?

Github Copilot Pro+ and Opencode or Claude Code Max?

@Afinetheorem They also acquired Harvest

Bending Spoons is so unusual. Buy every failed tech company you remember from a decade ago, fire most people, move leadership to Italy, (??), profit? I don't get it. This tiny Italian company now owns Eventbrite, Evernote, Meetup, Vimeo, Issuu, AOL. What am I missing?

@ondrejhanslik @TreciaKS Anyone who started web development 15-20 years ago can do both and much more 😉

@dpihac @TreciaKS Nah, neither can do the other thing. It's not about being able to do it but about having the knowledge and experience to do it. In my experience any developer can read both FE and BE code, maybe even making small updates but not writing it from scratch.

Hot take Frontend devs can do backend but backend devs can't do frontend

@devjibare @TreciaKS There are more languages than Javascript, especially on the backend 😉

@dpihac @TreciaKS No way, you can’t even deal with CSS, but a front end dev who understands JavaScript and typescript well, can deal with any backend language

@imbktan Auto-generated spaghetti code is responsibility and liability of each individual contributor. How they use AI is up to them. Using rules, conventions and especially skills, it's possible to generate high quality code. Coding by hand will become unsustainable soon.

Our engineer flat-out refuses to use AI IDEs like Cursor or Claude Code. He will use LLMs to look things up or proofread, but he will not let them touch the actual code. That is basically turtle speed compared to using an AI IDE, so yeah, we ship slower. But at least I can sleep at night knowing the codebase will not slowly turn into auto-generated spaghetti that nobody really understands and that becomes expensive technical debt later.

@arvidkahl I tweaked Ralph loop overnight to add support for GitHub Copilot CLI (others probably work too, but haven’t tested), PRD creation and some nice UI. Then I wake up and see this 😅

If the Ralph guy hadn't rugged his shitcoin earlier, THIS would have done the very same. I personally love the speed at which these things change. Ralph isn't even 2 weeks old (anywhere outside SV), and it's already being integrated, step by step.

I have modified Ralph Wiggum loop overnight to: - support multiple providers - run autonomously with some guard rails - created basic UI and tweaked to my needs Best of all, this approach could also support LLMs used via API, not only CLIs.