My First Week With GPT-6 Astra
GPT-6 Astra was my main Codex driver for the last week, and I am back on GPT-5.5. That sounds harsher than my actual opinion…
Beitrag lesenBlog
Erfahrungen aus echten Projekten
Ich schreibe über Software, KI-gestützte Entwicklung, Laravel, SaaS, SEO, Tooling und alles, was sichtbar wird, wenn Software im Alltag genutzt wird. Meist technisch, aber nicht auf ein Thema beschränkt.
73k Leser im letzten Jahr
310+ Beiträge
Laravel, KI, SaaS, SEO
Neuester Beitrag
Neue Beiträge
Die neuesten Beiträge aus dem Blog.
Hetzner’s Cheap Cloud Tier Is Unavailable After Two Price Increases
Hetzner used to be the obvious cheap cloud option in my mind. On September 7, 2026, I checked its public catalog and found every…
Beitrag lesenCan Provider Routing Change LLM Outputs?
Provider routing can change an LLM's output when a request reaches a different model version, fallback model, parameter configuration, precision level, inference engine, region,…
Beitrag lesenHow Floating-Point Determinism Affects LLM Reproducibility
Large language model (LLM) inference is the process of producing output tokens from an input prompt. It uses floating-point arithmetic for matrix multiplication, attention,…
Beitrag lesenHow to Test a Nondeterministic LLM Application
Test an LLM application as a workflow that must meet behavioral requirements across repeated runs, not as a function that returns one exact string.…
Beitrag lesenCan Tokenizer Changes Affect LLM Output?
Yes. A tokenizer converts text into a sequence of integer identifiers called token IDs before a large language model (LLM) processes it. A token…
Beitrag lesenIs the Same Prompt Always the Same LLM Input?
No. The same visible prompt is not always the same input received by a large language model (LLM). The text in a chat box…
Beitrag lesenThemen
Direkt zu den Bereichen, über die ich öfter schreibe.
LLM
107
Laravel
71
Software Development
49
Claude Code
17
Guides
16
OpenAI Codex
13
AI Engineering
13
Business
12
Alle Kategorien
Alle Themen im Blog.
Guides: AI Engineering
13
API
4
Bootstrap
6
Business
12
Guides: Business
1
Claude Code
17
Comparisons
7
Cursor
2
DB
3
DevOps
3
Docker
5
Economics
12
Email
3
Email Marketing
1
Entrepreneurship
1
Freelance
8
Guides: Freelancer
2
Git
11
Grok
2
Guides
16
Inertia.js
3
JavaScript
2
jQuery
1
Laravel
71
Laravel Forge
1
Linux
5
LLM
107
Marketing
1
MySQL
1
News
8
Notes
1
Guides: Online Marketing
1
OOP
1
Open Source
5
OpenAI Codex
13
Guides: Outreach
1
Paddle
1
PHP
9
Productivity
7
Guides: Remote Work
1
SaaS
3
SEO
9
Guides: Social Media
1
Software Architecture
9
Software Development
49
TailwindCSS
2
Guides: Tech Stack
5
Guides: Tipps
1
VILT
1
Vue
8
Webdesign
1
WordPress
2
Neu nach Kategorie
Aktuelle Beiträge aus den Themen, über die ich öfter schreibe.
LLM
107 Beiträge- My First Week With GPT-6 Astra
- Switching from GPT-5.5 to GPT-5.6 Made Me Less Productive
- AI Is Now a Commodity
Laravel
71 Beiträge- Speeding Up Laravel Tests with tmpfs (MySQL in RAM)
- Laravel’s $touches Property: Automatic Parent Timestamp Updates
- Creating PDFs with Chart.js charts in Laravel using wkhtmltopdf
Software Development
49 Beiträge- Hetzner’s Cheap Cloud Tier Is Unavailable After Two Price Increases
- Switching from GPT-5.5 to GPT-5.6 Made Me Less Productive
- Codex CLI Has Responsive Tables in the Terminal
Claude Code
17 Beiträge- Capital One’s Security Agent Tries to Disprove Its Own Findings
- Kimi K3 Is Already Close to Claude on a Real Coding Task
- Anthropic Moved the Fable Deadline Again
Guides
16 Beiträge- How Floating-Point Determinism Affects LLM Reproducibility
- Can Provider Routing Change LLM Outputs?
- How Model Updates Break LLM Reproducibility
OpenAI Codex
13 Beiträge- My First Week With GPT-6 Astra
- Switching from GPT-5.5 to GPT-5.6 Made Me Less Productive
- GPT-5.6 Sol Uses Twice the Tokens of GPT-5.5
AI Engineering
13 Beiträge- How Floating-Point Determinism Affects LLM Reproducibility
- Can Provider Routing Change LLM Outputs?
- How Model Updates Break LLM Reproducibility
Business
12 Beiträge- AI Will Not Make Everyone More Productive
- Why I Do Not Recommend Fastmail for Primary Email
- AI in the Mining Industry
Alle Beiträge
Alle Beiträge im Blog, neue zuerst.
My First Week With GPT-6 Astra
GPT-6 Astra was my main Codex driver for the last week, and I am back on GPT-5.5. That sounds harsher than my actual opinion…
Beitrag lesenHetzner’s Cheap Cloud Tier Is Unavailable After Two Price Increases
Hetzner used to be the obvious cheap cloud option in my mind. On September 7, 2026, I checked its public catalog and found every…
Beitrag lesenCan Provider Routing Change LLM Outputs?
Provider routing can change an LLM's output when a request reaches a different model version, fallback model, parameter configuration, precision level, inference engine, region,…
Beitrag lesenHow Floating-Point Determinism Affects LLM Reproducibility
Large language model (LLM) inference is the process of producing output tokens from an input prompt. It uses floating-point arithmetic for matrix multiplication, attention,…
Beitrag lesenHow to Test a Nondeterministic LLM Application
Test an LLM application as a workflow that must meet behavioral requirements across repeated runs, not as a function that returns one exact string.…
Beitrag lesenCan Tokenizer Changes Affect LLM Output?
Yes. A tokenizer converts text into a sequence of integer identifiers called token IDs before a large language model (LLM) processes it. A token…
Beitrag lesenIs the Same Prompt Always the Same LLM Input?
No. The same visible prompt is not always the same input received by a large language model (LLM). The text in a chat box…
Beitrag lesenWhat Is a System Fingerprint in LLM APIs?
When an LLM API response includes a field such as system_fingerprint, it is identifying the provider’s backend configuration, not your prompt, account, device, or…
Beitrag lesenHow Model Updates Break LLM Reproducibility
A large language model (LLM) generates text, structured data, code, or tool calls from prompts. Hosted LLMs are not fixed functions of prompt text:…
Beitrag lesenDo Structured Outputs Make LLM Responses Deterministic?
No. For a large language model (LLM), structured outputs make responses more predictable in format, not necessarily in content or execution. A structured output…
Beitrag lesenHow Does Mixture-of-Experts Routing Affect LLM Repeatability?
Mixture-of-Experts (MoE) models split a large feed-forward network into smaller experts. A router selects only a few experts per token, improving efficiency but adding…
Beitrag lesenWhat Is Batch Invariance in LLM Inference?
Batch invariance means a request produces the same inference result when the server runs it alone, alongside other requests, at a different batch position,…
Beitrag lesen