DEV Community (@ThePracticalDev) header
DEV Community (@ThePracticalDev) profile picture

The Twitter account that launched https://www.teapawt.org/t.co/TJyCu2S5ZF. Built on @forem 🌱 On Bluesky @/https://www.teapawt.org/t.co/TJyCu2S5ZF No DMs — please email us for support!

Joined August 2014
Photo by DEV Community (@ThePracticalDev)Photo by DEV Community (@ThePracticalDev)Photo by DEV Community (@ThePracticalDev)Photo by DEV Community (@ThePracticalDev)Photo by DEV Community (@ThePracticalDev)Photo by DEV Community (@ThePracticalDev)Photo by DEV Community (@ThePracticalDev)Photo by DEV Community (@ThePracticalDev)Photo by DEV Community (@ThePracticalDev)Photo by DEV Community (@ThePracticalDev)
DEV Community retweeted
We’re back! 🎙️ Season 28 is all about building with AI: what happens when an agent deletes your database, why your app charged someone twice, and how to fix it. Hosted by @lolocoding, sponsored by @TimescaleDB. Episode 1 drops Dec 2! 🎧 codenewbie.org
2
9
2
12
2,031
Average request time inside Rails on DEV fell from about 70ms to 16ms in a month. Since October 19, autonomous agent loops have merged 351 PRs into Forem. How the loops are split, and every configuration that broke on the way. { author: @bendhalpern } dev.to/dailycontext/our-loop…
1
4
1,651
A 14-day gem cooldown guards against hijacked releases. So what happens when the only fix for 21 security advisories is brand new? @tech_christine walks through how two Rails apps patched the same day anyway. 🧵
1
1
2
2,493
The way past it is one command: bundle update rails --conservative --cooldown 0, which overrides the Gemfile for that run only. Skipping still means checking the advisory's source, the checksums and the changes, since an urgent security update is also a good disguise.
1
751
So why didn't Dependabot open a Rails PR? Its security updates skip the cooldown, but only for advisories GitHub has reviewed, and as of November 9 none of the 21 Rails advisories were in its database. dev.to/cseeman/when-to-skip-…
743
Robotics club kid, four years of a Development Planning degree, then a Master's in Computer Science. With junior roles feeling harder to get and AI changing the job, share your own path into tech, and whether you plan to stay. { author: @maame_codes } dev.to/maame-codes/how-did-y…
1
1
1,658
A typo in <component :is> can break a page, and the linter, compiler or TS check can miss it. @jacobandrewsky shares the typed wrapper that made Vue dynamic components safer. 🧵
1
4
4,199
Two problems with the default: :is accepts a plain string, so a typo like 'AsewomeButton' slips through, and the props passed in aren't typechecked, so a div can end up with a text prop and a click handler.
1
2
870
The fix is a DynamicComponent built on Vue's render function: string component names are rejected (native tags like 'div' still work) and props are typechecked. Usage stays almost the same, just <DynamicComponent> in place of <component>. dev.to/jacobandrewsky/how-to…
1
1
880
DEV Community retweeted
tomorrow 👀
2
3
13
2,899
New author alert! 🎉 @devxmay committed predictions to git before testing whether kernel-level eBPF evidence helps diagnose slow AI agents, then reports what held up and what didn't. First DEV post! dev.to/032_mayank_d7cf900fa0…
2
2
13
2,005
New author alert! 🎉 @shaarugesh28091 traces a black screen after sleep to a slow Bluetooth chip by reading the logs, then fixes it with one line of config. Resume dropped from about 6 seconds to 0.3. First DEV post! dev.to/sanjai_shaarugesh_145…
1
5
2,072
A chat box is a fine place for a conversation and a terrible instrument panel. With eight autonomous loops, three model subscriptions and dozens of open PRs, scrollback breaks down. So Forem's loops got a custom observatory. { author: @bendhalpern } dev.to/dailycontext/ask-more…
1
6
2,229
A 32,000-token chat on a medium-sized model can need 16 GB of memory just for its KV cache. @amananandrai walks through how LLMs learned to remember less, from attention basics to 3-bit caches, no PhD needed. 🧵
3
2
6
2,277
Then the system tricks: PagedAttention borrows an idea from 1960s operating systems and cuts the cache into small blocks, prefix caching shares one system prompt across users, and TurboQuant stores cached values in about 3 bits with no retraining.
1
2
802
If a coding agent already caught the obvious bugs, what is a human reviewer supposed to do? A review prompt to adapt for AGENTS.md, a short process log for PRs, and why reviewers shouldn't act as a proxy for AI feedback. { author: @grayside_sb } dev.to/googlecloud/what-are-…
1
3
11
2,772
One week left to enter the @PresageTech SmartSpectra SDK Challenge! ⏳ 🫀 Build an app that responds to real-time human signals captured through a camera 🫀 $3,000 in prizes across 10 winners 🫀 Submissions due Nov 22 dev.to/challenges/presage-20…
2
7
59
4,354