1. Home
  2. Detail

@lion.mr2: النهر الثالث للعراق. #capcut #fyp #movie #explore #parati

حسين
حسين
Open In TikTok:
Region: IQ
Wednesday 23 September 2026 15:09:24 GMT
8822
421
6
17

Music

Download

No Watermark .mp4 (0MB) No Watermark(HD) .mp4 (0MB) Watermark .mp4 (0MB) Music .mp3

Comments

user1953161608426
سادس زوم :
باع العراقيين قبل شنو جان جاذب انتباههم و هسة شنو التفاهة الي تجذبهم
2026-09-23 17:22:19
22
hr_7m0
ابن عجاج 🇮🇶 :
2026-09-23 17:27:47
3
To see more videos from user @lion.mr2, please go to the Tikwm homepage.

Other Videos
MISTAKE NO. 4/10 ✶ Ignoring my AI costs until the bill hit $700 🥲 A huge chunk of that bill came from paying for answers I already had… Amy uses AI to find nutrition information for the foods people log. Originally, every search went straight to a paid AI endpoint, specifically Perplexity Sonar or Gemini Flash, even if we had already searched for that exact food before. But people eat the same things constantly. The same coffee every morning. The same breakfast. The same few lunches and dinners on rotation. So I added a cache. Before making another paid AI request, Amy checks the database (Supabase) first. If we’ve already looked up that food, it reuses the nutrition data we saved instead of paying AI to find the same answer again. Around 60 to 70% of those requests can now be answered from the cache. Caching wasn’t the only optimization I made, so I can’t attribute the entire reduction to it. But it was a major reason my bill went from around $700 to $221 the following month, even with more users. My new rule: before paying AI to do something, check whether your app has already paid for the answer. Next mistake: putting sign-up at the beginning of onboarding when it should have been at the end 🫠 #softwaredeveloper #coding #appdevelopment
MISTAKE NO. 4/10 ✶ Ignoring my AI costs until the bill hit $700 🥲 A huge chunk of that bill came from paying for answers I already had… Amy uses AI to find nutrition information for the foods people log. Originally, every search went straight to a paid AI endpoint, specifically Perplexity Sonar or Gemini Flash, even if we had already searched for that exact food before. But people eat the same things constantly. The same coffee every morning. The same breakfast. The same few lunches and dinners on rotation. So I added a cache. Before making another paid AI request, Amy checks the database (Supabase) first. If we’ve already looked up that food, it reuses the nutrition data we saved instead of paying AI to find the same answer again. Around 60 to 70% of those requests can now be answered from the cache. Caching wasn’t the only optimization I made, so I can’t attribute the entire reduction to it. But it was a major reason my bill went from around $700 to $221 the following month, even with more users. My new rule: before paying AI to do something, check whether your app has already paid for the answer. Next mistake: putting sign-up at the beginning of onboarding when it should have been at the end 🫠 #softwaredeveloper #coding #appdevelopment
8 custom Starbucks drinks you have to try, order like a pro and impress everyone. Which one will you try first? #StarbucksHacks #SecretMenu #CoffeeLovers #StarbucksDrinks #DrinkRecipes
8 custom Starbucks drinks you have to try, order like a pro and impress everyone. Which one will you try first? #StarbucksHacks #SecretMenu #CoffeeLovers #StarbucksDrinks #DrinkRecipes
بستێک خاکی سلێمانی جێگام بوایە😍.. #slemani #gorani #fyp #actives? #forupage_trendingvideo
بستێک خاکی سلێمانی جێگام بوایە😍.. #slemani #gorani #fyp #actives? #forupage_trendingvideo
Drinking failure #drink #funny #fyp #funnyvideos😂 #fail
Drinking failure #drink #funny #fyp #funnyvideos😂 #fail


About

Legal