orthosmile

Running & Meditation for peace and healthy life.

[رياضة منزلية @SjVea]

تمرين منزلي خفيف وبدون إجهاد للمفاصل : بدون قفز وبدون جري
هذه الحركة تساعد في :
* تنشيط وتحريك عضلات الساقين والجزء السفلي من الجسم
* زيادة النشاط البدني وحرق بعض السعرات الحرارية عند الالتزام بها.
* تحسين اللياقة العامة والقدرة على الحركة بسهولة من المنزل.

ملاحظة: للحصول على أفضل نتيجة للجسم، احرص على دمج التمارين مع نظام غذائي صحي ومستمر.

[صحة و لياقة @History_qq]

قم برفع اللياقة البدنية لديك من خلال هذه التمارين البسيطة

[Masculine Path @masculinepath04]

At 70, you’ll look 30.

Do this asap👇

[Bryan Johnson @bryan_johnson]

In 51 days:

> 37% reduction in ground contact time
> from 267 ms to 167 ms
> 87% improvement in reactive strength
> 45% reduction in knee valgus under impact
> transitioned muscular braking to achilles tendon recoil

I'm 49 years old and I'm training to dunk for the first time in my life.

[Maekwansri @KwansriR]

มื้อเช้ากับการทำกินง่ายๆแต่สารอาหารต้องครบ
-ปลาแซลมอนนาบกระทะ
-สลัดถั่วเข็มกับเห็ดแชมปิญอง
-ไข่ต้ม
-อะโวคาโด
-ซาวโดว์

[โค้ชปกรณ์ สอนลดน้ำหนักด้วยวิธีคุมอาหาร @cosmetic_of_p]

https://t.co/Acb4XY2GSh

#ลดไขมัน #ลดน้ำหนัก #อาหารคลีน #ลดความอ้วน #ลดพุง

[Siam Blockchain @siamblockchain]

ก.ล.ต. ออกกฎใหม่ Travel Rule โอนคริปโตต้องระบุตัวตนผู้โอนผู้รับ สกัดฟอกเงิน

อ่านเพิ่มเติม: https://t.co/7rHokT33nN

[Maekwansri @KwansriR]

เต้าหู้อ่อนนึ่งทำกินเองได้ง่ายมากๆ
1.นึ่งเต้าหู้ลงบนน้ำเดือด ประมาณ5-7นาที
2.หั่นขึ้นฉ่ายกับลวกคะน้าเตรียมไว้
3.ผสมน้ำมันหอย ซีอิ๊วขาว น้ำตาล น้ำต้มสุก กระเทียมสับ งาขาวคั่ว คนให้เข้ากัน
4.จัดผักใส่จานวางเต้าหู้ลงไปแล้วกรีดเต้าหู้ให้ไม่ขาดจากกัน ราดน้ำซอส(ข้อ3) ลงไป ก็อร่อยแล้ว

[Injective 🥷 @injective]

Since its recent launch, Injective @USDC is quickly becoming the canonical standard across dApps and the Injective ecosystem.

Here are just some of the platforms that support Injective USDC today across spot, perpetual, lending, borrowing, and more markets. 🧵👇️

[Bryan Johnson @bryan_johnson]

Lower risk of dying from any cause by 11% in 8 weeks.

Protocol:
> 3 workouts a week for 8 weeks

For each workout:
> 4 min at 90-95% effort
> followed by 3 min easy
> repeat 4 times

Your improved cardiorespiratory fitness is the risk reduction.

[Jobescape @jobescape_ai]

Full Claude Course

Master Claude to automate repetitive tasks, build apps without coding, create real portfolio projects, and become more valuable at work.

[BT beartai @beartai]

ในที่สุดวันนี้ก็มาถึง ลายน้ำข้าวหลามตัดจาก Gemini เอาออกได้แล้วนะ !
.
สาวก Gemini ทราบกันดีว่า เวลาเจเนอเรตภาพจากโมเดล มักจะมีลายน้ำข้าวหลามตัดบาง ๆ อยู่ตรงมุมขวาล่างเพื่อเป็นการยืนยันว่ามันคือภาพที่เจนมาจาก AI ไม่ใช่ภาพออริจินัล ซึ่งสำหรับคนที่ไม่อยากให้มีลายน้ำอันนี้ เราเอาออกได้สักที หลังจากที่ผู้ใช้งานบ่นมาสักพัก
.
ฟีเจอร์ ‘Media watermark’

ฟีเจอร์ที่ใช้เอาลายน้ำข้าวหลามตัดออกจะอยู่มุมซ้าย กดเข้าไปจะขึ้นชื่อมาว่า ‘Media Watermark’ (สำหรับคนที่ใช้ภาษาไทยจะขึ้น��่า ‘ลายน้ำของสื่อ’) ให้ลาก Cursor ไปวางแล้วจะขึ้นตัวเลือกมาเป็น On, Off คือเปิดและปิดลายน้ำ หลักการก็ง่าย ๆ ถ้าจะเจเนอเรตภาพแล้วไม่อยากได้ลายนำแค่กดไปว่า Off ภาพที่สร้างขึ้นมาก็จะไม่มีลายน้ำให้เห็นแล้ว แต่ถ้าใครรู้สึกว่าต้องมีก็เปิดต่อไปได้ 
.
แต่ถึงลายน้ำที่เราสามารถมองเห็นได้จะหายไป แต่ไม่ได้หมายความว่าลายน้ำที่ให้คอมพิวเตอร์ตรวจจะหายไปนะ ก่อนหน้านี้ฝั่ง Anthropic เ��ิ่งลงนามฝังลายน้ำในเนื้อหาที่เจเนอเรตโดย Claude ไป หลักการเจนภาพจาก Gemini ก็ยังมีฟีเจอร์แบบเดียวกันอยู่ คือฝังลายน้ำที่ตาเปล่าเรามองไม่เห็นอย่างที่ Google เรียกว่า ‘SynthID’ และ ‘C2PA’ นั่นแหละ
.
ซึ่งอันนี้จะเป็นลายน้ำที่ฝังอยู่ในทุกงานที่เจนออกมาอยู่แล้ว ดังนั้นสำหรับคนที่กลัวปัญหา��่าจะแยกภาพจริงกับภาพ AI ไม่ออก ก็สามารถส่งภาพ หรือคลิปให้ AI ประจำใจเราอ่าน SynthID ก็จะรู้ได้ทันทีเลยว่าภาพถูกเจนมาหรือของจริง หรือถ้าใครไม่ได้ใช้ AI ก็สามารถเข้าเว็บไซต์ https://t.co/y1pAjAG99G แทนได้เหมือนกัน
.
อ่านบทความฉบับเต็มได้ที่ https://t.co/vQmtfmUYIQ
.
#BTbeartai #MediaWatermark #Gemini #ลายน้ำ

[Shushant Lakhyani @shushant_l]

Here's how to turn Claude into your reusable AI employee so you can finally stop typing the same prompt 47 times a week.

---

1. Claude Skills are reusable instruction packs, basically teaching Claude the job once instead of becoming its unpaid reminder department.

---

2. Create a Skill when you keep using the same prompt, workflow, documents, formatting, or checklist like it's Groundhog Day at work.

---

3. Every Skill revolves around a SKILL.md file that tells Claude what to do, when to do it, and how not to freestyle into another dimension.

---

4. Add templates, scripts, PDFs, spreadsheets, examples, images, or other resources so Claude has receipts instead of vibes.

---

5. Claude figures out what you're asking, finds the relevant Skill, loads it, follows the instructions, and gets to work like it actually read the onboarding manual.

---

6. Prompts are one-time instructions, while Skills are reusable workflows for people who have officially retired from copy-paste athletics.

---

7. Memory remembers things about you, while Skills execute predefined workflows, because remembering your preferences isn't the same as doing your job.

---

8. MCP connects Claude to external tools and services, while Skills tell Claude how the actual workflow should work without causing workplace chaos.

---

9. Keep each Skill focused, specific, modular, tested, and packed with examples because "just figure it out bro" is not documentation.

---

10. Build Skills for writing, coding, research, marketing, design, data analysis, productivity, and basically anything you repeatedly make Claude do.

---

To learn more, check the infographic.

---

Get all my visual guides for free here: https://t.co/p3DBZe1GLQ

---

[Shushant Lakhyani @shushant_l]

Here's how to build Agentic AI that actually does the work instead of just confidently telling you how to do it.

---

1. Give your agent a goal, because "figure something out" is how both AI and Monday mornings go terribly wrong.

---

2. Let it reason through the problem, preferably before it starts pressing digital buttons everywhere.

---

3. Add planning so your agent breaks giant missions into tasks instead of having an existential crisis.

---

4. Give it memory so every conversation doesn't feel like meeting a very talented goldfish.

---

5. Connect tools so it can search, email, code, use databases, and finally earn its electricity bill.

---

6. Let it execute actions, because an agent that only talks is basically a consultant with unlimited tokens.

---

7. Add feedback so it can notice when its brilliant master plan was actually nonsense.

---

8. Install guardrails unless you enjoy giving autonomous software unrestricted access and hoping for character development.

---

9. The basic agent loop is Goal → Plan → Act → Observe → Reflect → Replan → Complete, aka AI overthinking professionally.

---

10. Level 0 is a chatbot, while Level 5 is a strategic agent that might start asking why you're still employed.

---

11. Single agents are great for focused jobs where creating an entire AI Avengers team would be slightly dramatic.

---

12. Multi-agent systems split work between specialists because apparently AI has discovered corporate departments too.

---

13. Planner + Executor is the classic combo: one makes the plan, the other discovers why the plan was optimistic.

---

14. Researcher + Writer works beautifully when you want facts first and confident keyboard smashing second.

---

15. The Critic Pattern reviews outputs, basically giving your AI another AI whose entire job is saying "Hmm, not quite."

---

16. Human-in-the-Loop keeps important decisions human-approved, which is comforting for everyone involved.

---

17. MCP connects agents with tools, databases, and apps so your AI can finally leave the chatbox and touch grass digitally.

---

18. Track accuracy, task success, cost, latency, tool efficiency, and human intervention before declaring your agent "revolutionary."

---

19. Watch for hallucinations, infinite loops, wrong tools, lost context, and permission errors, aka the Agentic AI blooper reel.

---

20. The formula is simple: Goal + Reasoning + Planning + Memory + Tools + Actions + Feedback + Guardrails = AI that actually gets stuff done.

---

To learn more, check the infographic.

---

Get all my visual guides for free here: https://t.co/p3DBZe1GLQ

---

[Shushant Lakhyani @shushant_l]

Here's how to build your own personal AI agent, so you can finally delegate work to someone who never says, "I'll do it tomorrow."

---

1. Pick one job for your AI agent first, because giving it 47 responsibilities on Day 1 is basically corporate onboarding.

---

2. Choose the AI model carefully, because this is literally the brain you're hiring without an interview.

---

3. Write a strong system prompt so your agent knows its job instead of confidently freelancing its own career path.

---

4. Give it clear goals, rules, tone, and output formats before it develops its own personality disorder.

---

5. Add memory so it remembers what happened five minutes ago better than half your meetings do.

---

6. Connect your knowledge so it can actually read your documents instead of spiritually guessing what's inside them.

---

7. Give it tools like web search, email, calendars, calculators, and code so it can graduate from talking to doing.

---

8. Connect your apps so your agent can work across Gmail, Slack, Notion, GitHub, and everything else you forgot you subscribed to.

---

9. Add planning so it can break big goals into smaller tasks instead of staring at them like we do on Monday mornings.

---

10. Build workflows that turn "do this thing" into an actual sequence of actions instead of another motivational quote.

---

11. Use understand → plan → execute → review → improve, otherwise you've basically built an intern with Wi-Fi.

---

12. Add human approval for important actions because autonomous does not mean "send whatever you want, buddy."

---

13. Add permission checks because your AI agent absolutely does not need access to your entire digital existence.

---

14. Log important actions so when something explodes, you can investigate instead of blaming "the algorithm."

---

15. Measure accuracy, speed, task completion, tool success, and failures because vibes are unfortunately not an analytics framework.

---

16. Start with a single agent because building twelve agents before one works is just creating an AI management problem.

---

17. Keep prompts simple because your agent shouldn't need a PhD to understand "summarize this PDF."

---

18. Test frequently because "it worked perfectly once" is the unofficial slogan of broken automation.

---

19. Improve its prompts, memory, knowledge, tools, and workflows continuously as you discover where your digital employee gets confused.

---

20. Start small, make it reliable, then expand, because the best AI agent isn't the fanciest one, it's the one that actually finishes the job.

---

To learn more, check the infographic.

---

Get all my visual guides for free here: https://t.co/p3DBZe1GLQ

---

[C.L.A.U.D.I.A @DaAcervo]

Normalmente, todos consideram brotos de brócolis saudáveis, até você ler as pesquisas sobre NRF2 que saem da Johns Hopkins, daí o negocio vai pra outro nivel!

O sulforafano é o ativador natural mais poderoso da via mestra de desintoxicação do corpo já descoberto.
Então, deixa de parecer legume e começa a parecer farmacologia.
O NRF2 é o interruptor mestre que controla mais de 200 genes protetores simultaneamente.
Quando ativado, ele coloca em funcionamento toda a rede de defesa antioxidante do seu corpo de uma vez.
Não um gene. Duzentos.

O sulforafano ativa o NRF2 com mais potência do que qualquer outro composto natural já testado.
É isso. Uma molécula. Duzentos genes.
A relação entrada-saída mais eficiente da ciência nutricional.

• Induz as enzimas de desintoxicação de fase 2 que neutralizam carcinógenos antes que eles danifiquem o DNA.
Estudado de forma aprofundada na Johns Hopkins há mais de 25 anos.
• Reduz o tamanho e o número de tumores em modelos animais de câncer.
Ensaios clínicos em humanos estão em andamento ativamente.
• Atravessa a barreira hematoencefálica. Neuroprotetor contra marcadores de Parkinson em pesquisas preliminares.
• Melhora os marcadores do transtorno do espectro autista em um ensaio controlado randomizado.
Um dos poucos compostos que demonstraram isso.
• Reduz os marcadores inflamatórios em todos os tecidos testados, incluindo intestino, fígado e cérebro.
• Fortalece a capacidade do seu corpo de processar e eliminar toxinas ambientais, metais pesados e partículas de poluição atmosférica.

Os brotos contêm de 10 a 100 vezes mais precursor de sulforafano do que o brócolis do supermercado.
E a maioria dos suplementos de sulforafano não tem mirosinase, a enzima necessária para converter o precursor no composto ativo.
Sem mirosinase, a cápsula é biologicamente inerte.

Cultive o brócolis em casa.
Pode ser em vasos.
Compre sementes para germinar.
De água de 3 a 5 dias desde plantar a semente o broto comestível.
É barato.
Coma de 30 a 60 g por dia.
Ou compre um suplemento que contenha explicitamente mirosinase (verifique o rótulo, a maioria não tem).

O composto quimioprotetor mais poderoso da natureza cresce em um vaso na sua bancada de cozinha por menos do que o preço de um café.

[HealthyBody @HealthyBody321]

Hip Strength Moves💯🔥

[Karl Matt Button │ Apex Fitness Adv. @KarlApexFit]

10 years ago I herniated a disc in my back.

- Pain
- Sciatica
- Inability to train

But now I deadlift over 220kgs.

Here are the 4 Simple Exercises I used to get my back strong and pain free:

= Thread =

[Mobility Exercises | Pain Relief Workouts @FeelGoodMovez]

how to fix rounded shoulders at home

[กายภาพตัวอ้วน (chubbyphysio) @chubbyphysio]

6 ท่านี้ คนที่นั่งทั้งวัน
-สะโพกติด,เส้นประสาทแนวกระเบนเหน็บตึง เพราะโดนนั่งทับนาน

เราช่วยเขาขยับหน่อย ช่วยกล้ามเนื้อหน้าสะโพกยืด ช่วยทำให้ข้อสะโพกไม่ติดมากกนะค้าบบ

#ตัวอ้วนชวนฟิต