1. Home
  2. Detail

@dayal.mandal7:

Dayal Mandal
Dayal Mandal
Open In TikTok:
Region: SA
Saturday 07 November 2026 15:20:51 GMT
0
0
0
0

Music

Download

No Watermark .mp4 (2.38MB) Watermark .mp4 (2.38MB) Music .mp3

Comments

There are no more comments for this video.
To see more videos from user @dayal.mandal7, please go to the Tikwm homepage.

Other Videos

Why long AI chats cost more (and the trick that cuts it by 90%) Why do long AI chats cost more? The longer your chat with ChatGPT, Claude or any AI model gets, the more every single message costs, even a tiny one. Here's why, and how prompt caching cuts it by 90%. How AI chats really work: 1. AI models don't remember your conversation. Every time you send a message, the whole conversation is sent again: your first message, its answer, your second, its answer, plus the new line at the bottom. 2. The model reads all of it from the top, then answers. Message 1 is cheap. Message 20 carries the 19 before it. That's why a long chat gets slower, and more expensive when you pay through the API. 3. You pay in tokens, small pieces of words. Input tokens are everything the model reads; output tokens are what it writes. On Claude Sonnet, reading costs $2 per million tokens and writing costs $10. 4. Example: paste a long report and ask for a short summary. The model reads around 10,000 tokens and writes about 500. That's around 2.5 cents, and most of that cost is reading. The trick: prompt caching. Most of what the model reads is the same every time: your instructions, your documents, the early part of the chat. You mark that part and the API keeps it ready. The next time, that part costs 90% less, ten times cheaper to read. Saving it the first time costs about a quarter more, so it pays off from the second use. Two catches: - The cached part has to match word for word from the very top. Change one word near the start and you pay full price again. Keep the stable stuff first and the changing stuff last. - The cache only lasts a few minutes by default, so it works best when the same thing gets read again and again. Three habits to save money on AI: 1. Start a fresh chat when the topic changes. 2. Put your fixed instructions first. 3. Building with the API? Turn on prompt caching. Useful for anyone using ChatGPT or Claude every day, and anyone building AI apps, AI agents or AI automations who wants lower API costs. Step-by-step lessons that teach AI properly, from the very beginning to building real systems, are inside my community. If you're interested, the link is in my profile. #aiagents #claude #aiemployee #aiautomation
Why long AI chats cost more (and the trick that cuts it by 90%) Why do long AI chats cost more? The longer your chat with ChatGPT, Claude or any AI model gets, the more every single message costs, even a tiny one. Here's why, and how prompt caching cuts it by 90%. How AI chats really work: 1. AI models don't remember your conversation. Every time you send a message, the whole conversation is sent again: your first message, its answer, your second, its answer, plus the new line at the bottom. 2. The model reads all of it from the top, then answers. Message 1 is cheap. Message 20 carries the 19 before it. That's why a long chat gets slower, and more expensive when you pay through the API. 3. You pay in tokens, small pieces of words. Input tokens are everything the model reads; output tokens are what it writes. On Claude Sonnet, reading costs $2 per million tokens and writing costs $10. 4. Example: paste a long report and ask for a short summary. The model reads around 10,000 tokens and writes about 500. That's around 2.5 cents, and most of that cost is reading. The trick: prompt caching. Most of what the model reads is the same every time: your instructions, your documents, the early part of the chat. You mark that part and the API keeps it ready. The next time, that part costs 90% less, ten times cheaper to read. Saving it the first time costs about a quarter more, so it pays off from the second use. Two catches: - The cached part has to match word for word from the very top. Change one word near the start and you pay full price again. Keep the stable stuff first and the changing stuff last. - The cache only lasts a few minutes by default, so it works best when the same thing gets read again and again. Three habits to save money on AI: 1. Start a fresh chat when the topic changes. 2. Put your fixed instructions first. 3. Building with the API? Turn on prompt caching. Useful for anyone using ChatGPT or Claude every day, and anyone building AI apps, AI agents or AI automations who wants lower API costs. Step-by-step lessons that teach AI properly, from the very beginning to building real systems, are inside my community. If you're interested, the link is in my profile. #aiagents #claude #aiemployee #aiautomation
mệt mà vui 🤭🤭#xuhuongtiktokk
mệt mà vui 🤭🤭#xuhuongtiktokk






Replying to @Valentina Riccobon Estas galletas craqueladas de brownie son los más rico y chocolatoso que hay! Y lo mejor es que son muy fáciles de hacer! ingredientes: - 180 gr de chocolate de buena calidad - 110 gr de mantequilla - 50 gr de cacao - 1 cda de café instantáneo - 200 gr de azúcar - 2 huevos fríos - 1 cdta de esencia vainilla - 1 cdta de polvos de hornear - 100 gr de harina - 1/4 cdta sal - 60 gr de chips de chocolate Se van a un horno precalentado a 180 grados por 8-10 min. Las mías quedaron perfectas a los 9 min. *Es normal que se inflen un poco en el horno y después se bajen, no se asusten 😊 *Es importante que los huevos estén recién salidos del refri y de hornear la masa rápido, para que se forme el efecto craquelado en la parte de encima. Cuéntenme si las hacen 😍 #cookies #brownie #chocolate #galletas #galletasdebrownie #chile #foryou #Receta #recetasfacilesyrapidas
Replying to @Valentina Riccobon Estas galletas craqueladas de brownie son los más rico y chocolatoso que hay! Y lo mejor es que son muy fáciles de hacer! ingredientes: - 180 gr de chocolate de buena calidad - 110 gr de mantequilla - 50 gr de cacao - 1 cda de café instantáneo - 200 gr de azúcar - 2 huevos fríos - 1 cdta de esencia vainilla - 1 cdta de polvos de hornear - 100 gr de harina - 1/4 cdta sal - 60 gr de chips de chocolate Se van a un horno precalentado a 180 grados por 8-10 min. Las mías quedaron perfectas a los 9 min. *Es normal que se inflen un poco en el horno y después se bajen, no se asusten 😊 *Es importante que los huevos estén recién salidos del refri y de hornear la masa rápido, para que se forme el efecto craquelado en la parte de encima. Cuéntenme si las hacen 😍 #cookies #brownie #chocolate #galletas #galletasdebrownie #chile #foryou #Receta #recetasfacilesyrapidas

About

Legal