Rajiv Shah | Explaining AI
Region: US
Sunday 12 April 2026 00:58:06 GMT
4311
203
6
12
Music
Download
Comments
Erik @ Bourbon Friday :
“That’s a lot of trust…” this part. It works until it doesn’t and Cursor uses multiple models that’s trusting all models to follow the skill. Also, uses up more tokens. Seems bad all around.
2026-04-21 13:51:38
0
Rajiv Shah | Explaining AI :
Check out the full conference keynotes in their YouTube live stream.
2026-04-12 01:51:36
2
kyrylowalker :
What's left for us? Trying to fix and repair all the mess that will result from these stupid integrations. You have to be a genius to shove an architecture everywhere that allows and hides errors 100% of the time
2026-04-12 10:49:31
0
user1239571678379 :
Why wouldn’t be AI good at design soon enough too😭😭 Coder are truly drinking poison to quench thirst. First it was easy lookup googling, after it was like line of code auto complete with copilot, gathering training data from proposed output to coder choice, after it was building block of code like function and class and gathering again improvement by coders adustments. After that it was full blown IDE integration with a "helper" and giving away data comparing the advice with implementation. After that whole agent workflow with planning, writing, testing, deployment, maintenance, update. Now you give the requirement and it build the whole thing end to end and you verify the job. What makes you think they won’t be able to design whole system soon enough too? The next coping mechanism in 2027 will be calling yourself "agent chief architect". Starting 2028 you’ll have a agent for a boss because he can manage more designers and agents than you and you’ll be dropped to IT superviser to do the Os update and be sure the Ethernet is plugged 💀
2026-04-12 03:17:18
1
TURFPTAx :
these are great updates for those in AI. keep up the great work!
2026-04-12 01:53:34
0
To see more videos from user @rajistics, please go to the Tikwm
homepage.
