CODERCOPS

We are group of people helping world to better program.

[CODERCOPS @codercops]

🚨 Code Red! The CODERCOPS are officially on duty. 👮‍♂️💻

We just launched on Product Hunt! 🎉

Drop a badge (or an upvote) and check us out here 👇
🔗 https://t.co/pyjxdVWBTW

#ProductHunt #BuildInPublic #IndieHacker #Coding

[Anurag Verma @anurag_629]

🚨 Calling all Web3 devs! New remote role on CODERCOPS.

A great team is hiring a Senior Smart Contract Engineer to build and scale secure dApps. ⛓️🚀

💻 Stack: Solidity, EVM, Smart Contract Security 🌍 Location: 100% Remote
Build your profile in one chat and apply instantly: 🔗 https://t.co/Dav2PKU0lc

#Web3Jobs #Solidity #SmartContracts #CryptoHiring #CoderCops

[CODERCOPS @codercops]

🚨 New remote role on CODERCOPS!

Blackbird Labs is hiring a Senior/Staff Backend Engineer to build the future of restaurant tech. 🍽️🚀

Build your profile in one chat and apply instantly: 🔗 https://t.co/TmLCt4fGK4

#Kotlin #Backend #RemoteJobs #TechCareers #CoderCops

[CODERCOPS @codercops]

Must apply!

[Anurag Verma @anurag_629]

🚀 New remote opportunity live on CODERCOPS!

We’re excited to share that Blackbird Labs, Inc. is looking for a Senior/Staff Backend Engineer to join their remote team. They are building a platform that connects restaurants with their guests through personalized rewards and their universal currency, $FLY.

What they are looking for:

🔹 5+ years of experience building and deploying backend APIs
🔹 Strong proficiency in Kotlin (or another statically typed JVM language)
🔹 Deep knowledge of RESTful APIs, Domain-Driven Design, and SQL
🔹 Passion for mentoring and fostering a strong engineering culture

⚡ Skip the long forms.

Create your CODERCOPS profile in just one chat and apply directly: 👉 https://t.co/9xkaBSRCK6

#BackendEngineering #Kotlin #RemoteJobs #TechHiring #SoftwareEngineering #CoderCops

[CODERCOPS @codercops]

Must apply

[Anurag Verma @anurag_629]

Hiring a Senior Web3 Full Stack Dev at Milepact! 🛠️

We’re building a trust-first freelance marketplace with escrow-backed agreements and need a strong builder to harden our early product.

💰 $50-$70/hr | Remote 💻 Solidity, Next.js, Foundry, viem/wagmi ✅ Paid test (150 EUR)

Apply below 👇

https://t.co/3yCpidm5DO

#Web3Jobs #SolidityJobs #Nextjs #CryptoJobs #Hiring

[Anurag Verma @anurag_629]

I've extra energy today 💪. What should I cook today 🤔🤠?

[Anurag Verma @anurag_629]

@heyblake https://t.co/o4b7Jni8Lb

[CODERCOPS @codercops]

@16vchq Hire cops https://t.co/6CtfAM4x5A

[Anurag Verma @anurag_629]

If you run a site behind Cloudflare, there's a config change worth making this week, not in September.

Cloudflare just split AI bot traffic into three buckets: Search, Agent, Training. Starting September 15, any crawler that doesn't clearly declare which bucket it's in gets blocked by default on pages that carry ads. That's the mechanism — but the actual decision you need to make is upstream of the deadline: which of these three do you actually want hitting your site, and does your current setup even distinguish them?

Here's what most teams have right now: one blanket robots.txt rule or one Cloudflare bot-fight toggle, applied to all three categories identically. That was fine when "AI crawler" meant one thing. It doesn't anymore — a Search crawler indexing your docs for an AI answer engine is a completely different tradeoff than a Training crawler hoovering your content for a future model's weights, and an Agent crawler acting on a user's behalf mid-session is different again.

Go set this explicitly instead of inheriting whatever the default becomes on September 15. Decide per-category, on purpose, before a platform decides it for you by default.

Have you actually looked at your bot traffic logs broken out by category, or is it still one bucket in your head?

#WebDev #Cloudflare #AI #DeveloperTools

[CODERCOPS @codercops]

most small businesses don't need a custom website. they need a working one. a few pages, a contact form, google maps, a whatsapp button, real seo, hosting that stays up. codercops builds and runs all of it for $19/mo. you never touch a cpanel login.

[CODERCOPS @codercops]

an AI agent with write access to your real data is a horror story waiting to happen. so ours writes nothing on its own. profile edits, job posts, starting a search: each one stops at a preview card you confirm first. the agent proposes, you decide what ships.

[CODERCOPS @codercops]

whoever is left, the founders interview. a real call, guided by a kit we build per candidate. the client gets the finalists and our actual notes, then runs the last round. no mystery box. they see the same work we did.

[CODERCOPS @codercops]

most "screening" is a resume skim and a vibes call. that is how you end up hiring a great talker instead of a great developer. here is the actual deep screen we run before anyone meets a client, step by step.

[CODERCOPS @codercops]

then the part people find strange. a second pass whose only job is to argue this person is the wrong fit and build the case against them. anyone who survives that is real signal. it is also where a few of our own favorites quietly drop out.

[CODERCOPS @codercops]

most internal tools die the day the project ships. you wrote it for one job, it does the job, nobody opens it again.

we open-source the ones that aren't tied to a client (ogcops, chatcops, no login) so they outlive the project. using them daily keeps them honest.

[CODERCOPS @codercops]

posting the role was never the hard part. reading the pile is. codercops hands you 3 to 5 finalists our founders already interviewed. you run the last round, and the success fee only bills if you hire one of them.

[CODERCOPS @codercops]

most "ai apps" pin one model and call it a day. our hiring chat picks per turn: easy stuff hits gpt-5 nano, harder turns step up to mini, then full. a failed turn auto-retries one tier up. cached input is 90% off, so you only pay for the smart model when the turn needs it.

[Anurag Verma @anurag_629]

Myth: a 10M-token context window kills retrieval, just dump everything in.

Reality:
- you pay per input token, so cost balloons
- latency scales with context size
- models still lose facts buried in the middle

Big context is a tool, not a dumpster. Retrieval keeps it small and sharp.