яσчαℓтℓя retweeted
We got Opus 4.6 and GPT 5.3-Codex on the same day. The AI usage cost today gonna be in direct negative correlation to the stock / crypto markets 📈 #ai #openai #anthropic
яσчαℓтℓя retweeted
Love to see Node kickin' ass like this. Node 22:
→ makes importing ESM into CJS files just work
→ added native file globbing
→ Introduces a task runner
Thank you, Bun?
яσчαℓтℓя retweeted
If anyone wants to stake and get rewarded BIG time here’s my blast.io referral links…
blast.io/PCBKG
blast.io/EQC9K
blast.io/J0KJQ
яσчαℓтℓя retweeted
#RunsOnRadix dApp Demo: CaviarNine 🖥️
Join @Kristus_maximus & @TronnOliver as they walk you through the essentials of the @CaviarNine platform.
Learn about staking, adding liquidity to LSU pools, executing swaps, & the seven-day deposit feature.
caviarnine.com/
яσчαℓтℓя retweeted
🍿 just made 👉 top50.surge.sh
if you're under 10: wtf, cancel netflix asap
if you're over 40: kudos
if you're 50/50: touch some grass
reply with results 👇
яσчαℓтℓя retweeted
Radix TVL has reached $25M in just eight weeks after the Babylon mainnet upgrade 🎉
defillama.com/chain/Radix
яσчαℓтℓя retweeted
Websites can now optimize themselves:
🔍 AI designs website variations & auto-launches A/B test
📊Results refine your next variations
🔄Process is automatically repeated when your results are stat sig
Basically Google Optimize on steroids
Try it: Keak.com
яσчαℓтℓя retweeted
Copying, dissecting and learning from @oguzyagizkara's beautiful design 🙇♂️ Please go and check his work
Insane level of detail, you can pick a single piece and spend a day analysing every small bit ✨
Original / My copy & variations
Some details ↓
яσчαℓтℓя retweeted
Very simplistic version of SEO 101 for indie hackers 👾
My websites have crossed a total of 200,000 monthly Google clicks -- achieved with $0 marketing budget.
I wrote this as a napkin SEO 101 for a friend, but then it struck me with "why not share it with everyone."
Research keywords:
Use @ahrefs or a similar tool to check the keywords you want to rank for, e.g. "case study." Type in "case study" first and then try to find more keywords. You should aim for the ones with the good ratio difficulty to volume (more volume, less difficulty). Long-tail keywords (i.e. consisting of a few words) are your friends usually.
Carry out on-page optimisations:
* H1 with your main keyword, like "case study." There can be only one H1 on page.
* H2, H3 mentioning keywords.
* <meta description> and <title> tags with keywords and good catchy descriptions -- they'll appear in Google search and affect your click-through-rate. Unleash the best copywriter in you.
* <img alt> with keywords and decent descriptions.
* Mention keywords you dug up in your KW research phase generously throughout your texts, but with no implications for the text readability or quality.
* Using slugs in the URL usually helps, i.e. sitename[.]com/case-study-with-gorgeous-mockups
* Overall, you can check your on-page SEO at https://pagespeed[.]web[.]dev
Use Google Search Console to monitor whether you're ranking, how you're ranking, & for what queries. It's free. You can submit a sitemap there, too. It's useful if you have multiple pages -- helps Google index your site faster.
Now, go out & try to get backlinks to your page. Mentioning your keywords in the <a> anchor helps. I.e. <a href="yoursite[.]com/your-page">Case Study XXX</a>.
You're on a hunt for do-follow links. <a href="yoursite[.]com" rel="follow"> is amazing, <a href="yoursite[.]com" rel="nofollow"> ain't bad, but it won't affect your domain rating that much.
Backlinks can be toxic, if pointing from a garbage site. Strive for those from reputable sources that have a good domain rating (DR).
Rule of thumb: Google understands whether you're reputable and should be shown to people by a myriad of factors, one of which is whether other reputable source point their fingers at you (in a good way). Kind of like in academia.
яσчαℓтℓя retweeted
Probably the best websites for 2023 🤯
→ chroniclehq.com
→ superlist.com
→ posthog.com
→ arc.net
→ bytes.dev
→ supabase.com
→ diagram.com
→ vercel.com
→ genius.design
→ liveblocks.io
→ tubecut.app
→ rive.app
→ resend.com
→ iconifyai.com
→ fantomshift.com
→ mindrudan.com
→ alignui.com
→ buildshowandtell.com
→ lazy.so
→ neon.tech
→ replay.io
→ ui.dev
→ 37signals.com
→ dub.co
→ shipfa.st
→ framer.university
→ read.cv
яσчαℓтℓя retweeted
v0.dev is 100% full-stack @vercel
◆ Built on @nextjs App Router
◆ AI by the @vercel `ai` SDK (sdk.vercel.ai)
◆ RSC & SSR Streaming on @vercel Functions
◆ Mutations via @nextjs Server Actions
◆ Data in @vercel KV (vercel.com/storage/kv)
◆ Flags on @vercel Edge Config
◆ Builds by @turborepo
◆ Social cards by @vercel OG Images
I encourage you to go and build your own AI-native products! – at San Francisco, CA
яσчαℓтℓя retweeted
Introducing v0 by @vercel - v0 generates custom components on the fly that you can copy and paste into your existing codebase.
Need a component? Just Ask - v0.dev
@jaredpalmer @shuding @maxleiter @mrncst @rauchg @cramforce @almonk and the team at Vercel
яσчαℓтℓя retweeted
My AI side projects this year:
◆ roomGPT.io: 2.1 million users
◆ restorePhotos.io: 400k users
◆ twitterbio.io: 200k visitors
◆ aicommits.com: 29k downloads
Here's exactly how I approach building AI side projects in 6 steps:
1. Think of an idea that excites me but that is very simple. I should be able to describe it in 5 words to anyone.
2. Take a weekend and block it off entirely to focus on building the app. I'll generally take 15-30 hours to build an initial version.
3. Try to incorporate the latest AI models so the end result is especially impressive (even though it's trivial with code, like 1 API call to an ML model)
4. Launch early then iterate based on feedback from users. The benefit of launching early is that you can pivot or work on a new idea if what you launched isn't working out so you don't sink too much time into a failed project.
5. Make it free + open source so folks can learn from it and are incentivized to share it and follow you. I gained about 20k Twitter followers and 18k GitHub stars from the 4 projects above.
6. Make sure the UI looks good and is straightforward to use. Nobody will use your project if it doesn't look great, so I typically focus a lot on this aspect.
I do want to call out that I do have a large twitter audience and that helps a lot. With that said, I've seen folks without any followers have successful side projects by doing the things I talked about, so I still stand by them!
яσчαℓтℓя retweeted
Problem: You want to display a dialog.
Common solution: Use state to track the dialog's visibility.
Alternative solution: Consider using a URL search parameter instead.
Benefits:
✅ Users can bookmark or share a link to the open dialog.
✅ If a dialog represents a unique location, it probably warrants a unique URL.
✅ You can use a plain anchor to open the dialog. Semantically, an <a> is the proper semantic tag for navigation. No button or onClick is required.
яσчαℓтℓя retweeted
📢 Yesterday, we released an experimental react-query-nextjs package for the app directory. Wrap your app in the <ReactQueryStreamedHydration> component to enjoy direct useQuery fetches on the server with suspense, which streams to the client 🎉
Example: tanstack.com/query/v5/docs/r…
яσчαℓтℓя retweeted
Comprehensive Rust
This is the Rust course used by the Android team at Google. It provides you the material to quickly teach Rust to everyone.
github.com/google/comprehens…
#rust #rustlang
яσчαℓтℓя retweeted
Replying to @housecor
It's in maintenance mode. I have been using `ts-unused-exports` + `typescript-eslint/no-unused-vars` rule for a while and works great. After some time we added it as a blocker on PRs.
The only pain point is it doesn't detect unused class methods or `import * as`
























![// Typical dialog state const [showDialog, setShowDialog] = useState(false); // Consider using the URL instead: site.com?showDialog=y](https://pbs.twimg.com/media/F2NHNzJXoAEfBPf.jpg?name=small&format=webp)




