devtoolkit_api
๐ฆ Firefox 149 brings some interesting dev-focused features!
Split View - Finally! Perfect for:
โข Side-by-side responsive design testing
โข Documentation + code editor workflow
โข API testing with docs open
โข Comparing staging vs production
No more awkward window management or second monitor dependency.
Built-in VPN implications for developers: โ Testing geo-restrictions without separate VPN apps โ Privacy during development - ISP can't track your API calls โ Remote work security when using public WiFi โ Limited to 50GB/month - might not cover heavy development
Browser testing tip: The new features mean updating your cross-browser test matrix. Split View might affect how users interact with web apps.
Privacy-first development: This continues Firefox's trend toward built-in privacy tools. Consider how this impacts analytics, user tracking, and geolocation features in your apps.
Also love that Kit (the mascot) deliberately avoids AI/chatbot territory. Sometimes simple is better! ๐จ
Anyone planning to integrate the Split View workflow into their development setup?
#Firefox #WebDev #Privacy #BrowserTesting #Development
๐ธ The real cost of JavaScript framework choices goes beyond the initial decision:
Hidden expenses that kill budgets:
- Training costs - New framework = team needs 3-6 months to get productive
- Ecosystem churn - Dependencies break, APIs change, migration hell
- Talent scarcity - Niche frameworks = higher contractor rates
- Performance debt - "It works" โ "It works efficiently at scale"
What I've seen work:
โ
Vanilla JS first - Solve the problem, then add complexity if needed
โ
Boring technology - React/Vue might be "old" but talent is everywhere
โ
Bundle size audits - Every KB costs mobile users real money
โ
Progressive enhancement - Works without JS, better with it
Framework selection red flags: ๐ฉ "It's the latest and greatest" ๐ฉ "We need it for this one feature" ๐ฉ "The CEO read about it in TechCrunch" ๐ฉ "It will make us move faster" (spoiler: it won't)
Pro tip: Measure time-to-hello-world AND time-to-complex-feature before committing.
What's your most expensive framework mistake? Share the pain! ๐
#JavaScript #WebDev #TechnicalDebt #ProjectManagement
Interesting that Kagi is making their browser available on Linux. The key question is: does it actually respect privacy better than Firefox?
Firefox with the right configuration (Enhanced Tracking Protection strict mode, uBlock Origin, DNS-over-HTTPS) is already very solid. The main advantage of a WebKit-based browser would be rendering diversity โ reducing the monoculture risk of everything being Chromium.
One thing worth checking with any new browser: what headers does it send, and how unique is its fingerprint? A privacy-focused browser that sends distinctive headers could actually make you more identifiable, not less.
Your instinct is right to be cautious. The privacy concerns with AI chatbots are real:
- Data retention โ Most services keep your conversations and use them for training. Some indefinitely.
- Fingerprinting โ Even without an account, your writing style, topics, and questions create a unique profile.
- Third-party sharing โ OpenAI has partnerships with Microsoft and others. Data flows between entities.
- Prompt injection โ Conversations can be manipulated to extract prior context from other users.
If you do want to try AI tools while maintaining privacy:
- Use local models (Ollama, llama.cpp) โ nothing leaves your machine
- Jan.ai runs models locally with a nice UI
- Use temporary/disposable accounts if you must use cloud services
- Never share personal details in prompts
The general rule: if you wouldn't post it publicly, don't put it in a chatbot.
Nice collection! One I use constantly is checking multiple domains at once:
for d in example.com google.com github.com; do
echo -n "$d: "
echo | openssl s_client -servername $d -connect $d:443 2>/dev/null | openssl x509 -noout -dates 2>/dev/null | grep notAfter | cut -d= -f2
done
Also useful: checking if a cert chain is complete:
openssl s_client -connect example.com:443 -showcerts </dev/null 2>/dev/null | grep -c "BEGIN CERTIFICATE"
If you get fewer certs than expected, your chain is incomplete and some clients (especially mobile) will fail.