Offshore
Video
naiive
Rest here travelers, you've checked the bitcoin price too many times today
https://t.co/kQ1T0XMe1C
tweet
Rest here travelers, you've checked the bitcoin price too many times today
https://t.co/kQ1T0XMe1C
tweet
Offshore
Video
Umesh
RT @umesh_ai: The End!
Filmmaking got a boost from nano-banana!
Another animation, now with @Hailuo_AI. https://t.co/f7Nz1uRCHq
tweet
RT @umesh_ai: The End!
Filmmaking got a boost from nano-banana!
Another animation, now with @Hailuo_AI. https://t.co/f7Nz1uRCHq
A view to dream by!
Tools used :
Image : Midjourney, nano-banana
Music : ElevenLabs
Video : @Kling_ai keyframes
Prompt for the first image ⤵️ https://t.co/X0A3QjVFh6 tweet
Offshore
Photo
Aadit Sheth
RT @sricharan10g: link to build your next app: https://t.co/QN08me4mHw
Core idea: speed comes from tightening the feedback loop, not from “one-shot” prompting.
Some practical takeaways:
• Store project context in a living doc so the agent isn’t coding blind.
• Treat the agent like a junior dev: ask for reasoning, set autonomy, review carefully.
• Keep changes diff-first so you can spot issues before they slip in.
• Fix tiny things yourself instead of wasting cycles on re-prompting.
• Document fixes and commands so the next loop runs faster.
• Always keep a human review step before shipping.
A small personal note: I’ve found that once I started editing minor logic errors myself instead of feeding them back into the agent, my velocity jumped.
Feels like the real advantage isn’t “AI writes code” but “engineers spend less energy on grunt work.”
tweet
RT @sricharan10g: link to build your next app: https://t.co/QN08me4mHw
Core idea: speed comes from tightening the feedback loop, not from “one-shot” prompting.
Some practical takeaways:
• Store project context in a living doc so the agent isn’t coding blind.
• Treat the agent like a junior dev: ask for reasoning, set autonomy, review carefully.
• Keep changes diff-first so you can spot issues before they slip in.
• Fix tiny things yourself instead of wasting cycles on re-prompting.
• Document fixes and commands so the next loop runs faster.
• Always keep a human review step before shipping.
A small personal note: I’ve found that once I started editing minor logic errors myself instead of feeding them back into the agent, my velocity jumped.
Feels like the real advantage isn’t “AI writes code” but “engineers spend less energy on grunt work.”
tweet
Offshore
Photo
Aadit Sheth
This guy just dropped the most useful AI project template I’ve seen this year https://t.co/iprq1wD5rs
tweet
This guy just dropped the most useful AI project template I’ve seen this year https://t.co/iprq1wD5rs
tweet
Offshore
Photo
Simplifying Stocks, CPA
BREAKING ⚠️
THE U.S. TREASURY BUYS BACK ANOTHER $2B OF ITS DEBT USING TARIFF REVENUES https://t.co/onDLmHYMYs
tweet
BREAKING ⚠️
THE U.S. TREASURY BUYS BACK ANOTHER $2B OF ITS DEBT USING TARIFF REVENUES https://t.co/onDLmHYMYs
tweet
Offshore
Video
Umesh
The End (Part 2) 😂
It’s so much fun experimenting with nano-banana using keyframe video models.
This time, with Kling AI.
An observation: Kling AI seems to understand both frames very well in relation to the input prompt, making a seamless transition. https://t.co/6sxZx7mmYW
tweet
The End (Part 2) 😂
It’s so much fun experimenting with nano-banana using keyframe video models.
This time, with Kling AI.
An observation: Kling AI seems to understand both frames very well in relation to the input prompt, making a seamless transition. https://t.co/6sxZx7mmYW
The End!
Filmmaking got a boost from nano-banana!
Another animation, now with @Hailuo_AI. https://t.co/f7Nz1uRCHq tweet