Pinned Tweet
Working on Mercury: clipboard and file sync across your LAN
Most tools that do this are full clipboard managers with history, search, pinning, cloud backup, and 200MB of RAM sitting there doing nothing
Mercury does two things: syncs what you copy between machines in under 500ms, and lets you send files with an accept/reject flow
That's it. Nothing else
AES-256-GCM encrypted in transit, passphrase never leaves your devices, discovery is mDNS so there's no server, no account, no config beyond a shared passphrase
Works across mac, linux, windows on the same network. Built with @wailsapp V3.
Still building but the core sync works 🪬
striker561.github.io/Mercury
Meet Mercury, an open-source desktop app for syncing your clipboard and transferring files between devices on the same network.
No cloud, no accounts, no subscriptions, just your devices and a shared passphrase.
GitHub: github.com/striker561/Mercur…
Wails v3.0.0 is here.
Build desktop apps with Go and your favourite web frontend.
Multiple windows. Native menus and system tray. Modular Go services. Generated TypeScript bindings. An in-app updater.
Watch the launch film. Release details below ↓
The user gets a DTO, the admin gets the full model, same repository, same query, different shapes
You can't accidentally leak admin data to a user endpoint if it never had access to it in the first place
Not an API, full Livewire but the principle holds whether you're returning JSON or not
BIGINT on exposed endpoints is just as bad. User 1847 already told me your platform is small.
ULID fixes the actual problem, sortable, opaque, no index chaos. And realistically anyone operating at the scale where this bites them already figured it out the hard way.
Or use BIGINT as your PK and expose any opaque ID publicly UUID, ULID, whatever.
The PK is what drives index performance, not the column you hand to users.
Nawa o!
See as person pikin just spray database columns with UUID everywhere na my first hint say na yeye ‘AI’ prompt dem use write this
UUID very horrible for Primary keys in high perf apps
@grok abi you too go prefer UUID to Bigserial or cheaply generated Numeric-only? 😏
Since I started using coolify years ago, mehn there is no going back 🙂↕️
6 Vercel alternatives to self-host:
1. Coolify(coolify.io/)
2. Dokploy(dokploy.com/)
3. Dokku(dokku.com/)
4. /dev/push(devpu.sh/)
5. Koyeb(koyeb.com/)
6. Siplane(sliplane.io/)
insignificant retweeted
Head of Design at Stripe on how to stop creating zombie UIs
If we only use AI to make the things that we already make, just faster, then we're definitely missing out on the most interesting part.
From my talk at @lennysan’s Summit, here are four ideas for raising the ceiling, not just the floor:
- Have a point of view [5:38]
- Encode your standards into the machine [9:07]
- Don't confuse done with good [12:21]
- Unleash creativity and artistry [16:19]
Full talk here: youtu.be/GLvFTMtw4Jk?si=-5Ts…
Here are my thoughts
Standard way to curb inflation is to increate price of things. But increasing the base price of essential everyday items only punishes new and low-level players. The goal is to lower or normalize overall income generation across the board while giving wealthy players big, optional reasons to burn their cash.
1. P2P Bank Tax: Add a 5–10% transfer fee when sending money to another player. Passive cash burn on every transaction.
2. Expensive Limited Edition Cosmetics: Banana Island mansions, rare vehicles, and high-end cosmetics that cost big money just for the status flex.
You can add events like Robbers (bots) that raid the houses if there are not enough securities, forcing rich players to protect your stash, you have to pay daily Estate Security / MOPOL dues. Or use the bank and pay Tax.
Once server-wide metrics show the economy has stabilized, the background spawn rate of robber events can automatically drop to a low, rare probability.
Investments should not drop huge returns as it used to as well, it should be normalized as well.
The inflation on #lagoslife is really something that I want to see how @Shalom_HeyEliy wisely solve.
I started asking on Reddit to get solutions from experts as I am not a gamer
Here is the thread
reddit.com/r/gamedev/comment…
“You already have the "no", if you don't ask. So, just ask”
This will stick with me
Replying to @pushpak1300
Back when I was still working at a consultancy, we ended up building a SaaS for sales people.
They had a background of literally going from door to door to try and sell something, so their rejection tolerance was really high.
One sentence they said always stuck with me since then: You already have the "no", if you don't ask. So, just ask.
insignificant retweeted
1/
LagosLife crossed 2M+ players within days of launch, with millions of visits daily and six-figure concurrent-user peaks.
And despite that completely unexpected growth, it has mostly remained responsive.
As a developer, I became more interested in how it survived the traffic than the game itself.
So I spent some time looking at the public clues around its infrastructure.
This is my attempt to reconstruct what may be happening behind LagosLife.
Important: this is observation and inference, not a confirmed teardown.
I don’t have access to their source code or infrastructure dashboard, and I could absolutely be wrong about parts of this.
But the clues are interesting. 🧵
2/
The strongest clue comes from LagosLife itself.
Their privacy policy says Cloudflare hosts both the game and its database.
Public fingerprinting also points towards React + Next.js.
That immediately changes how I think about the backend.
I don’t think we’re looking at the conventional architecture many of us would reach for:
Browser
→ Next.js
→ Node API
→ nginx
→ VPS
→ PostgreSQL
My suspicion is that LagosLife is far more Cloudflare-native.
Something closer to:
Browser
→ Cloudflare Edge
→ Workers
→ Cloudflare-managed storage/realtime primitives
Which is a very different scaling model.
3/
My rough reconstruction of the backend looks something like this:
Browser
↓
Cloudflare Edge
↓
Cloudflare Workers
↓
Then some combination of:
• D1 for persistent state
• Durable Objects for realtime state/presence
• KV or R2 for cache/assets/data
• Queues for asynchronous work
• external payment infrastructure
Only part of that is confirmed.
Cloudflare hosting is confirmed.
The specific services underneath are my inference based on the workload and the way Cloudflare’s platform is designed.
But the exact products matter less than the architectural principle I think LagosLife is using:
the backend probably does surprisingly little work per player.
4/
This is the part I think explains most of the scaling story:
LagosLife appears to be very client-heavy.
Think about the actual game loop.
Your character walking across the screen doesn’t need a database query.
Animations don’t need the backend.
Need bars counting down don’t need the backend.
UI interactions don’t need the backend.
Rendering the city doesn’t need the backend.
Timers don’t need the backend.
Cached world data doesn’t need the backend.
A huge amount of what makes the game feel alive can happen entirely inside your browser.
That means 100,000 people actively playing does not automatically translate into 100,000 people continuously hammering the backend.
"The full story" locked behind a ₦1,500 paywall. Paywalling an interesting interview in a community built on open source and shared context is … bad.
Respect to the team's effort and time, but micro-paywalls destroy the reach local tech breakdowns need. An email signup for a newsletter achieves your goal without killing reader interest. Right now, all this does is spike your bounce rate. I lost a bit of interest at the “sign up to read” then a paywall.
C'mon bro. We can’t build a solid ecosystem if every piece of shared knowledge is aggressively monetized.
Lagos Life was built overnight. Keeping it running was another story.
I spoke with @Shalom_HeyEliy about database migrations, fighting bots, the game economy and the engineering behind @lagoslifeapp.
The full story: africanengineer.com/i/lagos-…
Vectile works by indexing personal knowledge from several sources into an SQLite database with hybrid vector + full-text search and Reciprocal Rank Fusion, then lets you find things by meaning, not just by exact words.
Everything works on your machine locally. No API Keys.
Supports Obsidian vaults, folders of documents (Markdown, PDF, DOCX, HTML, TXT, CSV, JSON, YAML, XML, SQL, shell scripts, XLSX, PPTX, Jupyter notebooks, EPUB), Calibre libraries, and code repositories, including their commit history.
It can read scanned PDFs with OCR locally and you can install your own custom embedding model (.gguf)
You can also reach the indexed content with an AI agent using MCP
Imagine being able to find files and documents by searching for the meaning behind their content, rather than searching for text that doesn’t exist. That’s what Vectile solves
it's features can be exposed through MCP.
please give feedback❤️and stars✨
github.com/d3uceY/vectile


























