Most people think prompting is just writing better sentences Anthropic shows why that's wrong This 32-minute workshop breaks down how prompts are built for production Not prompt hacks Not "act as" Not another $300 course Actual systems Evals Edge cases Model migration Agent loops Good prompt is not one... message It's a workflow the model can follow, test and improve This matters because every new claude opus or sonnet will still reward the same rules Watch the prompting playbook Bookmark it before your next bad ai answer is just a bad setupshow more

Most people dont even know how to use computer and internet

hahah nowadays hear it… strange

agent loops handle edge cases perfectly

yesss

The real prompt is the workflow around it.

🤝

@0xwhrrari finally! someone addressing the real process behind prompts, not just hacky tricks.

🤝

@0xwhrrari what's up with the '10'? Curious!

evals and edge cases are where prompts become serious

yeeap yeap

production prompts need structure, not magic words

@asakura0x prompt ≠ magic word u goddamn right ser

model migration matters because setups should survive the next upgrade

its very important

Anthropic just dropped a real game plan for prodigy prompts

haha yeah

it's great that Antropic makes such helpful videos

its great also think so

saved it, Netflix is out - Anthropic tutorial instead

hehe

most bad prompts are really bad systems in disguise

agree

a good prompt is really a workflow the model can execute

good prompt - based thing

yep that’s wrong about prompting..

😼

been doing this wrong for a year 😪

👀👀👀

fr most prompt advice ignores everything that happens after the first response

😼

production evals are absolutely essential

😼

this distinction is more important than most people realize

提案することは、より良い文章を書くだけだと多くの人が考えていますが、Anthropicはそれが間違っていることを示しています。この32分のワークショップは、提案がどのように制作されるかを解説しています。提案のハックではありません。本当のシステム。評価、エッジケース、モデルの移行、エージェントループ。良い提案は1つのメッセージではありません。それはモデルが従うことができ、テストして改善することができるワークフローです。これは重要です、なぜなら、新しいクロード作品やソネットが、まだ同じルールを守って報われるからです。提案のプレイブックを見て、次の悪いAI回答が単なる悪い設定になる前にそれをブックマークしてください。

turkcesini yapip atıyom

@grok are the full videos on youtube?

i'm quoting this soo it stays on my timeline forever

Claude feels better for long reasoning. ChatGPT feels better for fast execution. The real win is not choosing one. It’s knowing when to use each.

Been Learning alot of prompts lately 💻 This explains a lot 👍🏻

Sounds insightful! It's great to see a deeper dive into prompting. Understanding systems really makes a difference in production. Can't wait to learn more!

Claude

the evals section is what separates prompting for demos from prompting for production. most tutorials skip it entirely. once you're running real workflows at scale you need to know when and how the model fails, not just when it works

This is an important shift in thinking. Prompting isn’t about clever wording, it’s about designing a reliable process the model can follow. Seeing more conversations around this on and lately.

Every new Claude model rewarding the same prompting rules is the part that makes investing in this skill worth it long term.

The “prompting = better sentences” myth is why people plateau. The mindset shift that helped me: stop writing prompts, start writing specs. A prompt you reuse and version is a system. A clever sentence you retype every time is a party trick.