Bilgin Ibryam ✓

@bibryam · joined 22 Jan 2009

Building agentic infrastructure. Writing about AI-assisted coding and agentic systems. Product manager @soloio_inc. I publish The Generative Programmer

84 821Followers

879Following

10 703Posts total

34.6KViews on collected posts

Derniers posts

Bilgin Ibryam ✓@bibryam22 Sep

Jev and LLMs: who does what? An LLM can write a reply. Jev can classify the request. Your code decides what happens next. A minimal example and eight early projects, from browser agents to context filtering. https://t.co/bgHxvfLqwZ https://t.co/2uLGop44Ao

1.1K views · 14 likes · 1 reposts · 7 replies Open on X →

Bilgin Ibryam ✓@bibryam21 Sep

Interesting MoE inference experiment: Treat MoE weights like JIT code. Per token: route experts → deduplicate the batch → run VRAM/RAM hits while NVMe misses stream → overlap I/O with compute → learn a hotter placement. COLIBRI: experimental, not production-ready. https://t.co

2.2K views · 11 likes · 2 reposts · 2 replies Open on X →

Bilgin Ibryam ✓@bibryam21 Sep

@bibryam this is actually a nice little workflow hack, reading the site first so the diagrams dont come out looking totally disconnected. followed u g 49 views · 1 likes · 1 reposts · 0 replies Open on X →

Bilgin Ibryam ✓@bibryam20 Sep

@bibryam Brand-matched diagrams in 60 seconds only hold if the skill locks type and palette. Freeform "make a nice chart" still drifts next run. 51 views · 0 likes · 0 reposts · 0 replies Open on X →

Bilgin Ibryam ✓@bibryam20 Sep

@bibryam The implication is that brand-aware diagram generation could make visual consistency a default property of content pipelines, not a separate design task. That also raises the value of maintaining a clean, machine-readable brand system. 40 views · 0 likes · 0 reposts · 0 replies Open on X →

Bilgin Ibryam ✓@bibryam20 Sep

🌟 Diagram Design Skill🌟 ⤷ A Claude Code / Codex skill for editorial-quality visual types, matched to your brand in 60 seconds by reading your website. https://t.co/aEyPHHCxRh https://t.co/AMeCKx9UTj

3.6K views · 50 likes · 13 reposts · 8 replies Open on X →

Bilgin Ibryam ✓@bibryam20 Sep

Turbovec is a local Rust vector index for RAG. It brings: • 2-bit or 4-bit compression • online inserts without a training phase • SIMD search on ARM and x86 • incremental, crash-safe saves v1.0 stabilizes its v7 on-disk format. https://t.co/Rzjgsp2NuR https://t.co/1OVEMgWIXq

2.7K views · 13 likes · 2 reposts · 1 replies Open on X →

Bilgin Ibryam ✓@bibryam7 Sep

@bibryam AI can generate the code. The harder part is keeping the reasoning, context and decisions behind it intact. Once agents build more of the system, that knowledge can’t stay scattered across files and people. This is where @ekosproject solves the problem. 4 views · 2 likes · 0 reposts · 0 replies Open on X →

Bilgin Ibryam ✓@bibryam6 Sep

@bibryam I'd add Working Effectively with Legacy Code. AI writes so fast that last week's generated service can already feel like a mystery. Feathers teaches you how to change code you didn't write without breaking everything around it. That skill just got a lot more valuable. 108 views · 2 likes · 0 reposts · 0 replies Open on X →

Bilgin Ibryam ✓@bibryam6 Sep

@bibryam One gap this list doesn't cover: if the same agent that wrote the code also flags what's questionable, it tends to confirm its own choices, not catch them. A separate reviewing role with no shared context catches a different class of miss than a human skimming the diff a 83 views · 0 likes · 0 reposts · 0 replies Open on X →

Bilgin Ibryam ✓@bibryam6 Sep

https://t.co/3Fd7GDv5bV 24.7K views · 334 likes · 34 reposts · 7 replies Open on X →

Face aux comptes de taille comparable

11 posts des 90 derniers jours, à côté de la tranche de 10K–100K abonnés. touche moins de monde que les comptes de taille comparable.

Vues médianes108ce compte935médiane pour 10K–100K

Portée, %0.13%ce compte3.66%médiane pour 10K–100K

Engagement, %1.52%ce compte1.52%médiane pour 10K–100K

Indicateur	Ce compte	Médiane pour 10K–100K	Rapport
Vues médianes par post	108	935	0.12×
Portée (vues ÷ abonnés)	0.13%	3.66%	0.03×
Taux d'engagement	1.52%	1.52%	1.00×

Autres comptes de cette tranche → Comparer avec un autre compte → Comment ces repères sont établis →

Growth & engagement

How the posts we collected actually performed: views and reaction rate post by post, what the audience did with them, and where the follower count goes.

Views per post

24.7K6 Sep

108

47 Sep

2.7K20 Sep

3.6K

4921 Sep

2.2K

1.1K22 Sep

Last 11 collected posts, oldest on the left. The scale is logarithmic: one post can outrun the rest a hundred times over.

Engagement rate per post

1.52%6 Sep

0.00%

1.85%

50.00%7 Sep

0.62%20 Sep

1.98%

0.00%

4.08%21 Sep

0.69%

2.05%22 Sep

Reactions — likes, reposts, replies and quotes — divided by views. Median for 10K–100K accounts is 1.52%.

What the audience does

Likes31.9%427 in total

Reposts4.0%53 in total

Replies1.9%25 in total

Quotes0.1%2 in total

Bookmarks62.1%832 in total

Share of every reaction we collected for this account. Replies mean argument, reposts mean endorsement, bookmarks mean the post was worth keeping.

The follower curve appears once this account has two daily snapshots — we take one a day, and this one is on its first.

Derniers posts

Face aux comptes de taille comparable

Growth & engagement

Views per post

Engagement rate per post

What the audience does

Comptes similaires