stuckinforloop

302 posts

stuckinforloop

@stuckinforloop

gamble with my life i’m here to bet some

somewhere Katılım Ekim 2019

728 Takip Edilen261 Takipçiler

stuckinforloop retweetledi

‏ً@omgsidewalks·23h

ADHD is one of the most painful things to live with. Not just because it's loud, because it's contradictory. You're capable of anything and motivated to do almost nothing. You understand everyone around you, but can't explain what's happening inside yourself. You have brilliant ideas, but no patience to finish a single one. You're a genius who can't handle an email, an extrovert who needs to be completely alone, a person full of advice who can't follow any of it. And the worst part? You know.

English

618

4.3K

116.9K

stuckinforloop@stuckinforloop·1d

yak shaving 🐂👀

stuckinforloop@stuckinforloop

it's that time of the month. I need to revamp my dotfiles 🏃‍♂️

English

stuckinforloop@stuckinforloop·1d

it's that time of the month. I need to revamp my dotfiles 🏃‍♂️

English

stuckinforloop@stuckinforloop·3d

I wish my full time job was to just read articles, brainstorm ideas, connect dots, share & learn perspectives, and move on to the next cycle not with AI Agents obv.

English

stuckinforloop@stuckinforloop·3d

anytime I see the word "load-bearing" in any content, I know it's written by Opus 4.7 the model just loves that word too much

English

stuckinforloop retweetledi

Dominik Tornow@DominikTornow·4d

Flaky tests break coding agents Agents interpret a flaky test failing as feedback to take corrective action. A subsequent pass reinforces potentially bad edits Reproducible tests that control non-determinism are the best foundation for your agent loop Blog post dropping soon

Dominik Tornow@DominikTornow

Unexpected: Go is a terrific language for deterministic simulation testing. The resulting ergonomics are delightful

English

4.3K

stuckinforloop retweetledi

Dilip Kumar@kmr_dilip·5d

Bro to bro: Start thinking of your future very seriously. Be overly curious, accept you know very little, be selfish to learn new stuff & don’t get into stupid office politics. The world is very cruel to people who are not doing new things everyday. You will become irrelevant.

English

753

6.1K

122.5K

stuckinforloop@stuckinforloop·4d

@arpit_bhayani welcome to AI pilled space 🫂

English

Arpit Bhayani@arpit_bhayani·4d

The way I test APIs looks so different now :) I make Claude write and run tests. I do not think I, or anyone, will ever write cURL or equivalent commands or use a tool in the near future. My workflow is now simple. Once all the core business logic is checked in and I am ready to test stuff, this is what I do: 1. make claude write tests 2. ensure it covers a good chunk of code 3. make claude write a test spec 4. spec has endpoints, responses, and potential errors 5. claude creates a testing-md file with key details 6. claude uses this file for every change, runs the relevant tests, and this keeps tests updated That is it. For me, claude code handles the mechanics - making HTTP requests, checking status codes, asserting on response shapes. I handle the judgment - whether the test cases actually cover what matters. The split works well because neither side is doing the other's job. (I just realised I am treating it like a co-worker). Given, claude can translate spec or even a thought into test code faster than I can set up any GUI tool, why do things manually :) By the way, the spec also doubles as documentation, which is a free side effect. Whether Claude uses pytest, go test, requests, httpx, or raw cURL under the hood does not matter to me. What matters is that the tests run, fail loudly when they should, and take minutes to write instead of an afternoon. I do not think I will open an API testing tool again. Writing the test logic was always the slow part. That problem is mostly gone now. I am getting AI-pilled, I guess :)

English

616

38.6K

stuckinforloop retweetledi

Arpit Bhayani@arpit_bhayani·6d

The more I build agentic systems at Razorpay, the more I understand that - at its core, it is an agentic loop with tool calls, integrations, and retrieval. The hard part is... actually making it run reliably, at scale, under real production load. And this is what makes system design even more important. Your AI system is still expected to scale. It will still need microservices, message queues, consistency guarantees, load balancing, work distribution, state management, rate limiting, throttling, fallbacks, service-to-service communication, QoS, etc. It is great that you are looking into AI and are interested. You should be. Everyone should be. But it is important not to skip system design and cs fundamentals. I know it seems overwhelming, but it is what it is. First principles are not going anywhere, and that is super essential for actually building applied AI systems and running them reliably at scale. If you are a backend engineer and are kind of skipping these things, pause and reflect once. It is always good to be great at system design, not because it will help you crack interviews (it will), but because it will make you meaningfully better at your job. Seeing it firsthand. Remember, you will not be shipping prototypes to production. The difference between prototype and production code is 15 components and 1000 commits.

English

1.5K

65K

stuckinforloop@stuckinforloop·5 May

@alex_whedon Congratulations!!!

English

stuckinforloop retweetledi

Alexander Whedon@alex_whedon·5 May

Introducing SubQ - a major breakthrough in LLM intelligence. It is the first model built on a fully sub-quadratic sparse-attention architecture (SSA), And the first frontier model with a 12 million token context window which is: - 52x faster than FlashAttention at 1MM tokens - Less than 5% the cost of Opus Transformer-based LLMs waste compute by processing every possible relationship between words (standard attention). Only a small fraction actually matter. @subquadratic finds and focuses only on the ones that do. That's nearly 1,000x less compute and a new way for LLMs to scale.

English

1.5K

2.9K

23.1K

12.6M

stuckinforloop@stuckinforloop·30 Nis

exactly!!! all these fancy setups that are showcased on tech twitter do not work when put into a production codebase and you can only do so much with compaction/sub-agents/etc.

Taelin@VictorTaelin

seriously, working with AI is MISERABLE for one and only one reason: having to re-explain the same thing "oh yeah this new session obviously doesn't know what proper case trees are, so let me explain it for the 5000th time in my life" I'm tired AGENTS.md doesn't solve this because it is impossible to fit the entire domain knowledge without nuking the context - it would be 1m+ tokens worth RAGs don't solve this, the agent won't search unknown unknowns SKILLs don't solve this unless I keep like a collection of 1750 skills with specific cuts of domain knowledge for each possible subset of my domain that I might need in a given chat, but that's a lot of manual work recursive LLMs or whatever don't solve this for the same reason, you can't dump a domain book and expect the AGENT will magically guess that it is supposed to search for a specific bit knowledge. unknown unknowns fine tuning doesn't solve this (OSS models suck and OpenAI / Anthropic gave up on user fine tuning) I honestly think a good product around fine tuning on your domain would be a major hit and an underdog lab should take this opportunity

English

stuckinforloop retweetledi

Adam@adamdotdev·28 Nis

I’m going through the craziest burnout I’ve experienced in my ~17 year career I’ve been sick for 16 days now, haven’t even been able to go for walks I kind of fucking hate AI I think all of these things are related

English

305

3.9K

308K

stuckinforloop@stuckinforloop·27 Nis

@ponnappa FYI the copy buttons on `join.realfast.ai` page after linkedin auth is not working

English

364

Sidu Ponnappa@ponnappa·27 Nis

we're starting to move our hiring funnel to an mcp frontend as a fizzbuzz and already 1 in 5 applicants don't make it through to a completed application but hey we used to reach offer stage with just ~1 in 1000 inbound applicants at gojek 10 years ago so not surprising i guess

sphinx@protosphinx

i’m actually surprised by the lack of ai enthusiasm and slow adoption in india. past tech waves had a much stronger curve. i’m not talking about enterprise adoption. they didn’t care about saas so it’s not like they’ll move to ai. i’m talking about dev exposure. if you ask 10 devs about claude code cli - 9 would have no clue. very few even care. that’s the surprising part. sign of times to come. bad omen ? yes.

English

8.4K

stuckinforloop@stuckinforloop·27 Nis

has anyone built an AI personal assistant based on Ari Gold's character from Entourage

English