tech

AI, Code & E-commerce

Articles on AI automation, web development, Next.js, SEO, and scaling e-commerce businesses. 120 articles.

I write about the tools and systems I actually build with — from multi-agent AI pipelines and MCP servers to Next.js full-stack apps, Supabase backends, and SEO automation. Every article comes from a real project: shipping production code, scaling e-commerce stores, or wiring up AI agents that do useful work. If you're a developer, founder, or just curious about how modern web stacks fit together, you'll find practical walkthroughs, honest trade-offs, and working code examples here. Topics span AI and LLM integration, cloud infrastructure (Vercel, AWS, Docker), developer tooling, and the business side of building software products.

All120 AI72 Automation25 Dev Tools18 Next.js16 SEO12 Web Development10 Engineering10 E-commerce5 Cloud5 DevOps4 OpenClaw4

120 articles

M4 Max vs M5 Max: Why I'm Skipping the 81,000 kr Upgrade

AppleMacBook ProM4 Max+4

Read article

M4 Max vs M5 Max: Why I'm Skipping the 81,000 kr Upgrade

My M4 Max upgrade verdict: $5,949 before U.S. sales tax versus 81,000 kr including Swedish VAT, with benchmarks and the new OLED MacBook Pro.

Uygar DuzgunOct 28, 2026

10 min read

Kimi K3.1 Pro: Early Reviews, API Access and Availability

AIKimi K3.1 ProAI Coding+2

Read article

Kimi K3.1 Pro: Early Reviews, API Access and Availability

Early Kimi K3.1 Pro reactions point to frontend and 3D work. Here is where to get API access, which plans qualify and what remains unverified.

Uygar DuzgunOct 23, 2026

6 min read

The LLM Black Box: What Goodfire's New Research Shows

AILLM InterpretabilityMechanistic Interpretability+2

Read article

The LLM Black Box: What Goodfire's New Research Shows

Goodfire says many LLM neurons are more legible than expected. Researchers disagree on the implications. What does that mean for AI builders now and later?

Uygar DuzgunOct 17, 2026

10 min read

Claude Haiku 5.5: Effort Levels and Practical Uses — Part 1

AIClaudeHaiku 5.5+3

Read article

Claude Haiku 5.5: Effort Levels and Practical Uses — Part 1

How I use Haiku 5.5 for translations, corrections, quick edits, and light deploy work, plus practical starting points for all five effort levels.

Uygar DuzgunOct 12, 2026

13 min read

My Long-Running Codex Goal: Four Days and Still Working

AICodexGPT-6.1 Sol+2

Read article

My Long-Running Codex Goal: Four Days and Still Working

My Codex goal passed four days. An honest account of GPT-6.1 Sol, an unreleased SEO product, repeated tests and what remains unfinished.

Uygar DuzgunOct 9, 2026

10 min read

Vape Website Performance: Inside Cigge's 99/100 Result

Web PerformanceEcommercePageSpeed+2

Read article

Vape Website Performance: Inside Cigge's 99/100 Result

I cut redundant browser work across Cigge’s storefront. The October 12 Vercel mobile snapshot shows 99/100, LCP 1.63 s and INP 128 ms—with checkout CLS still to fix.

Uygar DuzgunOct 2, 2026

13 min read

Gemini 4 Argon: Google's Comeback? Benchmarks and Pricing

TechAIGoogle+3

Read article

Gemini 4 Argon: Google's Comeback? Benchmarks and Pricing

Gemini 4 Argon has my attention. I compare benchmarks, introductory and later prices, and early reactions before testing Google's new model.

Uygar DuzgunOct 1, 2026

7 min read

OpenAI Dots vs Hermes: Is This OpenClaw 2.0?

OpenAIDotsHermes+4

Read article

OpenAI Dots vs Hermes: Is This OpenClaw 2.0?

I run Hermes for business automation. OpenAI Dots makes me curious: less maintenance, more integration, and a Sweden access catch.

Uygar DuzgunSep 29, 2026

8 min read

GPT-6.1 Sol vs Opus 5.5: Benchmarks, Price, Best Uses

TechAIOpenAI+2

Read article

GPT-6.1 Sol vs Opus 5.5: Benchmarks, Price, Best Uses

GPT-6.1 Sol promises more agent work for the budget. I compare its benchmarks, API price and early user reactions with Claude Opus 5.5.

Uygar DuzgunSep 29, 2026

7 min read

Claude Opus 5.5: Benchmarks, Efficiency, and What Users Notice

AnthropicClaudeAI Models+2

Read article

Claude Opus 5.5: Benchmarks, Efficiency, and What Users Notice

Anthropic says Opus 5.5 approaches Fable 5.1 on most tasks while costing less and running faster. Here is what the launch claims, early users, and the missing benchmark context tell us.

Uygar DuzgunSep 22, 2026

4 min read

OpenAI Adds Two More GPT-6 Models to ChatGPT and Codex

OpenAIGPT-6AI Models+2

Read article

OpenAI Adds Two More GPT-6 Models to ChatGPT and Codex

GPT-6 Sol and Luna have appeared in OpenAI's official model catalog. Here is what the model cards confirm, what users are reporting, and what still needs real testing.

Uygar DuzgunSep 22, 2026

4 min read

Jev Product Validation: 98/100 and a Confident Mistake

AITypeSafe JevE-commerce+2

Read article

Jev Product Validation: 98/100 and a Confident Mistake

Jev matched 98 of 100 product-copy checks, but approved a tenfold load error. Explore the English and Swedish results and confidence trade-offs.

Uygar DuzgunSep 19, 2026

11 min read

M4 Max vs M5 Max: Why I'm Skipping the 81,000 kr UpgradeMy M4 Max upgrade verdict: $5,949 before U.S. sales tax versus 81,000 kr including Swedish VAT, with benchmarks and the new OLED MacBook Pro.

Kimi K3.1 Pro: Early Reviews, API Access and AvailabilityEarly Kimi K3.1 Pro reactions point to frontend and 3D work. Here is where to get API access, which plans qualify and what remains unverified.

The LLM Black Box: What Goodfire's New Research ShowsGoodfire says many LLM neurons are more legible than expected. Researchers disagree on the implications. What does that mean for AI builders now and later?

Claude Haiku 5.5: Effort Levels and Practical Uses — Part 1How I use Haiku 5.5 for translations, corrections, quick edits, and light deploy work, plus practical starting points for all five effort levels.

My Long-Running Codex Goal: Four Days and Still WorkingMy Codex goal passed four days. An honest account of GPT-6.1 Sol, an unreleased SEO product, repeated tests and what remains unfinished.

Vape Website Performance: Inside Cigge's 99/100 ResultI cut redundant browser work across Cigge’s storefront. The October 12 Vercel mobile snapshot shows 99/100, LCP 1.63 s and INP 128 ms—with checkout CLS still to fix.

Gemini 4 Argon: Google's Comeback? Benchmarks and PricingGemini 4 Argon has my attention. I compare benchmarks, introductory and later prices, and early reactions before testing Google's new model.

OpenAI Dots vs Hermes: Is This OpenClaw 2.0?I run Hermes for business automation. OpenAI Dots makes me curious: less maintenance, more integration, and a Sweden access catch.

GPT-6.1 Sol vs Opus 5.5: Benchmarks, Price, Best UsesGPT-6.1 Sol promises more agent work for the budget. I compare its benchmarks, API price and early user reactions with Claude Opus 5.5.

Claude Opus 5.5: Benchmarks, Efficiency, and What Users NoticeAnthropic says Opus 5.5 approaches Fable 5.1 on most tasks while costing less and running faster. Here is what the launch claims, early users, and the missing benchmark context tell us.

OpenAI Adds Two More GPT-6 Models to ChatGPT and CodexGPT-6 Sol and Luna have appeared in OpenAI's official model catalog. Here is what the model cards confirm, what users are reporting, and what still needs real testing.

Jev Product Validation: 98/100 and a Confident MistakeJev matched 98 of 100 product-copy checks, but approved a tenfold load error. Explore the English and Swedish results and confidence trade-offs.