baranorhan

16 posts

baranorhan banner
baranorhan

baranorhan

@baranorhandev

AI & Data Engineer @ SF cybersecurity startup. Nights & weekends maker ➔ https://t.co/gYn2Dtlowy Creator of https://t.co/SaS0r00aCd

Ankara | SF Katılım Aralık 2012
103 Takip Edilen57 Takipçiler
Sabitlenmiş Tweet
baranorhan
baranorhan@baranorhandev·
I sit at a desk most of the day, and exercise instructions usually leave me wondering whether I understood the movement correctly. So I built a way to turn those instructions into something I can actually see. Posecode: human movement written as text and rendered in the browser.
English
1
0
1
5
baranorhan
baranorhan@baranorhandev·
@daronyondem claude codeun olsuturdugu memory ile codexin olusturdugu memory cakistiginda hangisi source of truth olarak kabul ediliyor
Türkçe
0
0
0
83
Daron Yondem
Daron Yondem@daronyondem·
Birkac aydir bu proje uzerine calisiyorum. Gunluk AI kullanimimin 90% bunun uzerinden geciyor. github.com/daronyondem/ag… Benim inanilmaz isime yariyor :) Belki size de yarar.
Türkçe
3
3
79
6.5K
baranorhan
baranorhan@baranorhandev·
Posecode is open source. Underneath the demo is a small movement language, a parser, range-of-motion checks and a Three.js renderer. Source: github.com/posecode-dev/p…
English
0
0
0
2
baranorhan
baranorhan@baranorhandev·
I sit at a desk most of the day, and exercise instructions usually leave me wondering whether I understood the movement correctly. So I built a way to turn those instructions into something I can actually see. Posecode: human movement written as text and rendered in the browser.
English
1
0
1
5
baranorhan
baranorhan@baranorhandev·
@arvidkahl same problem here maybe keep separate branches for the agents but let both use the same local services with isolated db schemas and index prefixes?
English
0
0
0
340
Arvid Kahl
Arvid Kahl@arvidkahl·
I want to run Clauded Code AND Codex on the same codebase at the same time, but I really don't want to have to deal with worktrees and provisioning db/redis/opensearch for each of them...
English
59
0
78
15.3K
baranorhan
baranorhan@baranorhandev·
@pentagoniac open weights help but the real question is whether the full stack can be audited training data evals system prompts deployment controls and compute still stay concentrated..
English
0
0
0
26
baranorhan
baranorhan@baranorhandev·
@srchvrs also better approach is to use multiple models as separate reviewers like one checks clarity and structure, one acts as a domain critic, one looks for unsupported claims and missing citations then compare the disagreements instead of accepting a single polished rewrite
English
1
0
0
26
Leo Boytsov
Leo Boytsov@srchvrs·
What's the motivation of people to not use LLMs to polish their writing and make it more understandable and idiomatic? So many poorly written papers. It's completely unforgivable in the era of ChatGPT when everyone can produce native speaker quality English prose. PS: I don't suggest direct generation of papers using LLM, only careful postediting and draft revision.
English
2
0
7
1.2K
baranorhan
baranorhan@baranorhandev·
@gdb been testing this kind of reasoning on posecode, a small language i’m building for turning movement descriptions into 3d animation the hard part is keeping joint constraints consistent across the whole motion curious how sol handles that (medium effort)
English
0
0
0
21
baranorhan
baranorhan@baranorhandev·
@arvidkahl this works when the tests and specs are actually good otherwise you’re just moving the code review problem somewhere else
English
0
0
0
22
Arvid Kahl
Arvid Kahl@arvidkahl·
Setting up the agentic harness to be a) test-first, b) documentation-centric, and c) configured by a detailed operation procedure spec makes not looking at the code possible. Without any of these three, chaos reigns. But with them, code review becomes optional most of the time.
antirez@antirez

It is my belief that many devs right now are not maximizing what they can do with automatic programming because they still look at the code. Doing it makes you the bottleneck. Your time is better invested in new ideas, QA, design, and asking yourself what is your goal.

English
32
5
65
8.2K
baranorhan
baranorhan@baranorhandev·
been building ai and data products for a few years but almost never shared the process trying to change that i’ll post what i’m building what breaks and what i learn
English
0
0
0
29

Keşfet

@daronyondem @arvidkahl @pentagoniac @srchvrs @gdb @elonmusk @BarackObama @taylorswift13