When its almost time for the quota reset and you still have some unused quota left.
Switch to the big gun models, love the instant difference it makes!
#vibecode#opencode#drafts
@RockstarGames How about adding building destruction physics? If the building holds missions, add a reconstruction time for it to get back to its original state and hold off the mission or reset the map on a new respawn. Do better!
#GTA6#GTAVI#RockstarGames
A plug and play fun little app, that can work with any launcher (for now pixel only, will do some changes for custom launchers) and its OPEN SOURCE.
Saw some folks dropping these in reddit but as an app or within a launcher, this one works universally.
https://t.co/HpvJlTQBzp
@ZeptoNow Your open box delivery is basically a scam to sell electronic devices that are packaged and in good physical condition but when its time for us to put it to work but realise that the gadget doesn’t even work. Its the customer’s loss?
Launching multiple products in the upcoming week!
- SeeMyBacklog
- GeldGrid
- Quicknotch
- Bento-Terminal
Which one pokes your curiosity? Will drop some teasers soon.
If you’re using electron to build your desktop apps even on this date, am sorry but you have to be from a non tech background trying to vibe code their front end product.
and to think that models like kimi k3.1 pro didn't even consider gemini part of its benchmark as it's competing with other heavy hitters, says a lot about gemini 😭
Although, gemini has successfully made Ecosystem with Ai than LLM standalone model
Sundar, Sundar, my friend, with respect, a benchmark table is not a comeback.
You invented the transformer. You had DeepMind, TPUs, YouTube, Search, Android, Gmail, and 3 billion Chrome users before anyone outside SF knew what an LLM was.
You had every advantage a company could have.
And yet in late 2026, when a developer opens a terminal, it isn’t Gemini. When a founder builds an agent, it isn’t Gemini. When my non-tech friends say “I asked the AI,” they don’t mean Gemini.
That’s the scoreboard that matters, and Google is losing it.
@GearsofWar's latest release on E-Day has to be the most pathetic state of a game coming from a AAA studio. An UE5 slop along with poor optimisation, game can't even stream texture into its pool even when it's in SSD.
Anyone looking into getting any “quote” from @livspace , spare yourself a trauma and simply don’t. Especially when they setup “pre-booking” camps, made a 25k pre-booking, decided not to go ahead, almost 5 years and still “processing my refund”.
@akadhanu and not to forget, the stealth models are pretty capable on their own and free while being able to do quite the heavy lifting with a good usage quota.
Switching between multiple Ai models? struggle with context eating half the token and compaction driving it all away or the prompt generated input output tokens are seriously high? Simply use headroom
You'll thank me later, if you haven't used it already
https://t.co/DM3HPDYuxv
Common apple L, flight’s at 7:25 to bengaluru, Left home only at 6, airports 30mins far. Car’s got my carplay paired and ready to go, meanwhile iPhone takes ages to handshake only to fail connecting. Disconnected -> paired again -> fails. Took the pixel out, instant connection!
Opencode's big pickle model is highly underrated, although stealthy. It seems to outperform gemini easily so far with the tasks I've given
#opencode#bigpickle
Subscribed to multiple LLM Providers, giving them all a shot to build my project at once. Interesting results
Gemini vs GLM vs Deepseek vs Qwen vs Grok and more.
#ai#agentic#productdev