Your next 10,000 developer signups start here. We build marketing campaigns so devs/agents try, adopt, and recommend your devtool without friction
Joined March 2016
- Tweets6.8K
- Following290
- Followers3.1K
- Likes4.7K
Hackmamba retweeted
I haven’t touched Claude Code or Codex CLI in a while.
The terminal era is over imo. It's the wrong interface for coding agents. Tabs are ephemeral, but context is persistent, and managing 30 tabs is pure cognitive overhead.
I don’t really need an IDE like Cursor either. I rarely navigate the whole codebase anymore.
The new primitive is the agent, not the file.
(Codex desktop app is the best agentic UI for now. But we’re still early.)
Hackmamba retweeted
There’s always so much context behind a product that users never get to see.
Boki’s Expert Insights feature brings the people closest to the work into the content process.
For this demo, Hemanth, one of the engineers behind the Boki Agent, shared what looked simple to users but was actually difficult to get right.
His insight was captured in Boki and turned into separate LinkedIn and X posts.
Here’s the process 👇
Hackmamba retweeted
I’m joining @hackmamba !!!
As Senior DevRel, I’ll handle socials strategy and help companies reach developers through content and community.
We have helped 50+ devtool teams ship product launches, create technical content, run paid ads, and build communities.
If you’re a dev tool or tech company looking for help, reach out in DMs.
It’s been a long time coming, and this was the best team I could ever ask for.
Excited to get things going!
Hackmamba retweeted
Let's talk about recall!
You learned that RAG retrieval can fail even when the right document is in your index. Here's what's actually going on.
Your HNSW search walks a graph of vectors but stops early if its candidate list is too small.
That limit is controlled by ef_search. Too low and it never reaches your document, it just hands back the closest thing it found before quitting.
Bump ef_search and it walks further. Your recall goes up.
See our YouTube channel for full breakdown.
Click here: youtube.com/watch?v=1-oV5g0R…
Hackmamba retweeted
Teams don’t usually have a distribution problem because they lack content.
The problem is that distribution gets treated like something to figure out after the work is done.
Write it. Publish it. Then ask: “Okay, how do we get people to see this?”
A better workflow starts earlier.
Hackmamba retweeted
Attackers use AI to find vulnerabilities faster. Your defenses should too.
Introducing CodeRabbit Security.
Map out your repository, hunt vulnerabilities in production code, verify each against real code, and ship fixes as PRs.
No speculation. No noise.
Hackmamba retweeted
We've raised a $143M Series C at a $1.5B valuation.
In 2023 we spent our time arguing that AI code needed its own reviewer. That argument is over. The work now is moving into the change itself.
This is where we think Agentic Change Management comes in.
Hackmamba retweeted
not me getting genuinely excited over an open source repo but here we are
@hackmamba dropped Creator Skills, seven Claude skills made for tech writers, and the deslop reviewer is quietly the MVP of the set. catches AI sounding writing before your readers do
repo is right here if you want to try it: github.com/hackmamba-io/crea…
Hackmamba retweeted
me reading "$10,000 grand prize" and suddenly remembering I know how to code 😎
Have you heard of @actiandev's The Heist?
30 days, 4 jobs to pick from, teams of up to four, and $10,000 on the table.
Build something that survives constrained hardware, the edge, or fully disconnected environments.
More details and RSVP here: luma.com/zj89lhf2
Hackmamba retweeted
Are you ready for the heist?! 🎮
You’ve got four jobs to choose from and 30 days to pull it off.
Your build has to keep running on-prem, at the edge, constrained hardware, fully disconnected.
You bring the plan. Actian VectorAI DB is your inside connection.
$10,000 is sitting there. One grand prize and no runner-ups.
September 4 to October 4. Your crew, your Discord, your track guide are all in the first comment.
Don’t leave the money on the table.
luma.com/zj89lhf2
Hackmamba retweeted
honestly one of the more fun build challenges I have seen in a while.
drop a reference image, build it live with Bit and Hope AI, submit before August 15, and you are in the running for $500.
no reason not to try this one
If you’ve been waiting for a reason to build and ship something, here you go! 🤝
The Bit Build Program gives you one month, one challenge, and a chance to win $500.
For Month 1, you’ll build a live product from a single reference image.
See all details below.
🧵
Hackmamba retweeted
If you write, market, or grow content in tech, this is worth checking out.
AI is changing how people discover information, and understanding why it cites some sources while ignoring others is becoming essential. Boki has brought together some brilliant minds in content and growth for this conversation tomorrow.
Your article ranks on Google. Then someone asks ChatGPT the same question, and it cites your competitor instead.
Tomorrow, Boki brings together some of the sharpest minds in content and growth to unpack why AI trusts some content and ignores the rest.
If you write, market, or grow content online, you won’t want to miss this one.
Join us live 👇
meet.google.com/jct-jrax-hgt
RSVP on Discord:
discord.gg/kzrDbqUa?event=15…
Hackmamba retweeted
one thing i like checking every week 😅 they just refreshed the job board with 20 new roles.
if you're in devrel, marketing, technical writing, or aiming for a leadership role, you might find something here.
direct link: devtooljobs.hackmamba.io/?pa…
Hackmamba retweeted
A DashboardState class holds the watchlist, prices, positions and P&L.
A @rx.event(background=True) task streams price updates on an interval. Every component bound to that state re-renders on its own.
That's a live trading dashboard in pure Python, no JS, and no manual refresh.
Computed vars handle the P&L math, rx.cond drives the alert.
Repo + full build here: reflex.dev/blog/how-to-build…
Hackmamba retweeted
I built a RAG pipeline that retrieves live web data before generating answers.
The goal was simple.
I wanted a pipeline that could work with information that changes, instead of relying on a static collection of documents.
Here's how I approached it. 🧵
Hackmamba retweeted
The one trait you should never lose
open.substack.com/pub/ichulo…
Hackmamba retweeted
Has anyone running a swarm of coding agents locally hit the runtime collision problem yet? Worktrees isolate files, but they all share one database and one set of ports.
Found an open-source take on isolating the runtime per branch, curious if others solved it differently
ito.ai/blog/moo-worktree-iso…
Hackmamba retweeted
Everyone repeats that GitHub's free runners are 2-core.
That's only true for private repos.
Public repos get 4-core Linux runners now. It changes the parallelism math: on 2 cores, Playwright's undefined worker setting resolves to a single worker, so you're serial by default.
On 4 cores, you get 2 out of the box, with room to push higher.
Worth knowing before you tune workers against advice written for a machine you're not running on.
endform.dev/blog/playwright-…
Hackmamba retweeted
Playwright CI feels slow, so you shard it. That's usually the wrong first move.
Sharding costs you CI minutes across several runners, a merge-reports step, and a shard count you have to rebalance as the suite grows.
Before any of that: cache the browsers, turn on fullyParallel, lift the CI worker cap, and run Chromium only on PRs. That holds a 100-test suite under five minutes on a single runner.
Shard when your numbers cross the line. Not when the run first feels slow.
endform.dev/blog/playwright-…
