Sabitlenmiş Tweet
I've been building a dev agent at Datadog to fix your prod errors automatically.
Demo linked from reply.
English
Dan Benamy
585 posts
@MichaelKuz I rename terminal tabs/windows or use desktop app or web which have nice session management
English
I keep losing track of what claude code session I'm using for what, and end up tabbing between a dozen terminal windows, reading and scrolling through the chats, until I find the right one. Is there something simple I'm missing?
English
@richardartoul Does it work to paste it into a human brain's context window?
English
if you’re worried about agi just paste a few MiBs of server logs into your context box and watch it freak out
English
Dan Benamy retweetledi
If you want LLMs to show charts/#dataviz of live data instead of text alone, now they can via the @datadoghq x @cursor_ai MCP app!
Cursor@cursor_ai
Use natural language to query logs, metrics, traces, and dashboards with the Datadog plugin.
English
Dan Benamy retweetledi
The trend towards cloud agents is directionally correct. I’ve worked via long lived ssh sessions for a decade, it’s great and even better with AI agents churning away async
I can close my laptop and walk away, they’re still working. When I reconnect, all the context is there.
English
Dan Benamy retweetledi
free alpha: early drop of a first party datadog cli
github.com/DataDog/pup
English
@felixge Feedback loops are the way.
Have you tried codex? Or multi-gen and pick the best?
English
However, I still spend a few hours convincing claude to not edit existing test cases. And I'm still not sure if I'm happy with the solution for some of the gnarly event ordering problems I encountered. There is a cacophony of invariants involved.
English
Dropped a new Go proposal for writing execution traces. Together with the existing reader, this allows interesting data filtering to reduce data volumes.
This is also the first time I'm trying to heavily lean on AI for an upstream Go contribution (1/N)
github.com/golang/go/issu…
English
Dan Benamy retweetledi
Dan Benamy retweetledi
build life systems in 2026. people over index on will power as it's perceived to be an admirable characteristic. will power is fickle. it fluctuates with stress, fatigue and hunger.
nearly half of your daily decisions are autonomous already (you don't think about them). you want to create if-then rules. if 7am, work out. if 5 pm, no more food for the day.
every choice you make taxes your mental energies. you want to make as few decisions per day as possible. saving your mental energy for things that are higher value like professional achievement or relationships. not on the mundane things like what and when to eat or exercise.
English
Dan Benamy retweetledi
In retrospect, the articles mocking Dario’s prediction that 90% of code would be written by AI by September seem to be very misguided. He seems to have been only off by a couple months (if that).
Boris Cherny@bcherny
@YashGouravKar1 Correct. In the last thirty days, 100% of my contributions to Claude Code were written by Claude Code
English
Dan Benamy retweetledi
This feels like a watershed moment. When the best engineers in the world are no longer writing a single line of code, and barely even looking at the code that's being generated.
This would have sounded insane just a year ago.
Hard to fathom what happens when the stuff @bcherny @steipete @karpathy are doing today becomes the norm for the rest of the software engineering field over the next few years.
Boris Cherny@bcherny
@YashGouravKar1 Correct. In the last thirty days, 100% of my contributions to Claude Code were written by Claude Code
English
Dan Benamy retweetledi
The lack of "feature design" is why so many products over time feel hollow or messy. This isn't visual design. This isn't architectural design. I thought that a short video lecture of what feature design is and a real case study of applying it in Ghostty would be helpful.
Feature design is the planning step behind how you're going to solve one or more user problems with a product feature: what that feature looks like, how it feels, and not just how its going to tactically solve these specific problems, but how that solution is going to interface with the edges of other features that currently exist or are planned to exist in the future.
English
I don't know what the KPIs are, but every time I open Github on mobile, it's worse than the last time.
Honestly, give me the textarea-like editor of 5 years ago any time.
20% of my screen wasted on line numbers, and sluggish as hell.
Jorge Manrubia@jorgemanru
We recently added paid monthly subscriptions to Fizzy (yes, we launched without them in place). It was the first time both 37signals and I used Stripe. My impressions as a developer were fantastic. Here’s the full implementation for anyone who wants to take a look. It includes the fantastic design and copy by @jasonzimdars. github.com/basecamp/fizzy…
English
Dan Benamy retweetledi
Bits AI SRE is now generally available for all customers, bringing always-on AI investigations to speed incident resolution and reduce risk.
Read all about it in @TheAIJournal1: bit.ly/4pfXDIR
English
I have two dev machines.
On my primary machine, I'm doing my regular work.
On the second, I have several coding agents working on my side projects. It feels like I have two interns working on a number of ideas I had but never been able kick off until now.
English
Dan Benamy retweetledi
At @datadoghq we've just published a blog with insights on #ruby performance based on surveying profiler data from thousands of real-world services: datadoghq.com/blog/ruby-perf…
English
Dan Benamy retweetledi
I keep thinking about this post and every time when I think about founders who are there it’s like: you have money and aren’t running out, you can build your dreams
Do you know how many people would trade places with you? So so so many people.
Touch grass. Build what you want.
Rohit Mittal@rohitdotmittal
most venture-backed companies end up in the same place: raised 3-4M seed (or 10-15M series A) built real product got to 300-500K ARR (or 2-3M ARR) VCs stopped caring founders stuck with high pref stack, can't raise more, can't exit, can't even shut down cleanly these aren't bad businesses they're just wrong cap tables the companies could throw off 500K-1M in profit annually hire 2-3 more people and grow to 2-3M ARR actually build something valuable but they're trapped in venture math that doesn't work anymore the discourse only covers two outcomes: unicorn or failure nobody talks about the third path: decent business, wrong structure thousands of quality founders stuck running companies VCs wrote off not because they're bad operators because they picked wrong market size or wrong cofounder or raised too early
English
I genuinely need someone to help me understand YT premium because I think it's for teens.
YouTube is my number 1 "streaming service", but the premium features don't land at all.
Zach Warunek@ZachWarunek
It’s odd how many people have a visceral reaction to the thought of YouTube Premium, but it’s quite literally the best subscription on the internet
English
Dan Benamy retweetledi
We built Datadog’s natural language querying features (variant of text->SQL) using a fine-tuned model, replacing prompted OpenAI models.
We did this explicitly for latency and cost purposes: the feature actually translates as you type in the UI, which required both <500ms latency, and would have been wasteful to do on a pay-per-token model like the hosted providers. We run it on our own pay-per-hour GPUs, allowing real time translation. Any UX that’s trying to feel like a tab-completion model (fast, user accept/reject) likely would benefit from similar approaches
#natural-language-queries" target="_blank" rel="nofollow noopener">docs.datadoghq.com/logs/explorer/…
English
I think 2026 will be the year I get an EV. Tax credits are expiring, yet manufacturers realize there's no consumer demand to go up, so they're cutting prices
English