私もPro 5倍プラン無料で受け取りました。 GPT6 Astraと一緒に使うと週間限度という言葉が無色に一日しかできず逆体感がひどいですが、それでも非常食がひとつ生じたようで心強いです。ありがとう @OpenAI
4
795
Thanks for @OpenAI and @AnthropicAI. Got accepted into the Codex and Claude Code OSS Programs, so that 6 months the the 5x/20x plans are free.
I'll come true SDD (Spec Driven Development) with these free tokens through compiler-powered evidence graph.
github.com/wrtnlabs/evidence
1
3
203
TypeScript v7 compiler toolchain:
- reduce 90% AI tokens by graph
- plugin support like typia and nestia
- integrated linter into ts compiler
github.com/samchon/ttsc
1
5
481
I made TS Compiler Graph MCP, that can reduce 90% token usage in both Codex and Claude Code.
With strong CoT compliance harness, it beats every code graph MCP tools, even in the general-purpose usage.
dev.to/samchon/i-made-ts-com…
1
5
50
24,078
From now on, TS plugins no more need specify tsconfig.json's compilerOptions.plugins. Just install, and just use.l with ttsc.
github.com/samchon/ttsc
5
159
@ryoppippi Your "@ryoppippi/unplugin-typia" has been revived as "@ttsc/unplugin".
Now any TS transformer plugin (typia, nestia, or your own) can plug into Vite, Webpack, Rollup, esbuild, and more — with zero extra config.
github.com/samchon/ttsc
1
4
12
1,292
I'm doing VR coding in AI coding era, to monitoring 5+ agents at once.
dev.to/samchon/vr-coding-for…
1
2
2,447
Local LLM Benchmark about Backend Generation.
GLM vs Qwen vs DeepSeek
dev.to/samchon/autobe-benchm…
1
2
215
Transformer and executor for TS v7 (typescript-go).
It's 10x faster than "ts-node", and can do type checking that "tsx" can't do. Also, it supports transformer libraries like "typia".
github.com/samchon/ttsc
2
2
10
3,280
Succeeded to build backend application from scratch with Qwen 3.5 27b with AutoBe. One thing amazing is, no quality difference between Claude Sonnet 4.6.
dev.to/samchon/autobe-qwen-3…
1
5
447
As a coding agent developer, this article was written after reviewing the source code of Claude Code, a top-tier project in this world.
dev.to/samchon/autobe-vs-cla…
2
5
204
I play VR boxing games on the company rooftop every day.
A clean hit for the first time in a while.
2
150
Finally, AutoBe achived 100% compilation success rate on small sized model "qwen3.5-27b". The more important thing is, difference of AI model does not affect to result quality.
This is the function calling harness.
github.com/wrtnlabs/autobe
2
2
8
2,882
CoT becomes 100% structurally guaranteed. Streaming + property order validation catches the AI skipping steps in real time, mid-generation. DeepPartial<T> fills corrections progressively as output streams in.
If you can verify the process, you can harness any domain.
3
703
Next subject: function calling for domains where the final answer is unverifiable.
Legal rulings, business decisions: you can't verify the conclusion. But you can verify the process. Model every mandatory reasoning step as typed schema properties, put the verdict last.
1
5
909
dev.to/samchon/qwen-meetup-f…
Same article on dev to blog for someone using Galaxy Fold like device suffering from layout broken. I will fix it ASAP.
3
16
1,886
I was invited by the Qwen team (@Alibaba_Qwen) to give a presentation at a Meetup. It is the story of the Function Calling Harness, which turns a 6.75% success rate into 100%.
autobe.dev/blog/function-cal…
14
63
526
39,942
@Ron github.com/nonara/ts-patch/p…
Let's support TypeScript v6 from TS-PATCH
1
2
606