Braindrop | Synister (IN)

Patching morphing resonating Synth sounds from my new VITAL preset pack ! @matttytel braindropsynister.gumroad.com/l/vitalpsypack #Plugins #Techno #Psytrance #Electronicmusic

@isro Jai Hind 🇮🇳

Chandrayaan-3 Mission: The image captured by the Landing Imager Camera after the landing. It shows a portion of Chandrayaan-3's landing site. Seen also is a leg and its accompanying shadow. Chandrayaan-3 chose a relatively flat region on the lunar surface 🙂 #Chandrayaan_3 #Ch3

Tipping feature will be a game changer for the creators on 𝕏. Imagine being able to tip a user directly from a post by simply tapping the tip icon. 💸

As a Uae raised 2nd gen person , I must admit the guy is an incredible visionary ! Even though i have my gripes with this place , it is technically my hometown and I miss it at times!

Elon Musk confirmed that "Calling" feature is coming soon on 𝕏 across all major platforms i.e. iOS, Android, Windows, Linux & Mac.

Cmmon champ .. let’s go AJ 🥊

Last one was solid!

Boxing schedule looks great! 🔥 Joshua vs Helenius Navarrete vs Valdez Usyk vs Dubois Canelo vs Charlo Fury vs Ngannou Edwards vs Rodriguez Prograis vs Haney Beterbiev vs Smith Eubank vs Smith 2 Zhang vs Joyce 2 Wood vs Warrington Catteral vs Linares Cameron vs Taylor 2

Oleksandr Usyk with an unorthodox approach to the speed bag in training for Daniel Dubois on Aug 26th… [🎥 @Usykaa]

Working on new stuff with @Surachai

@TateNews_ Without doubt ! @Cobratate I’ll Dj for you too 🥳🥳 in return for some entrepreneurial tips 😃✌🏽

@elonmusk 🤣🤣✌🏽

I'm calling the Myth of Context Length: Don't get too excited by claims of 1M or even 1B context tokens. You know what, LSTMs already achieve infinite context length 25 yrs ago! What truly matters is how well the model actually uses the context. It's easy to make seemingly wild claims, but much harder to solve real problems better. I highly recommend "Lost in the Middle: How Language Models Use Long Contexts" from Stanford. It's jam-packed with rigorous experiments that put popular long-context models to test. Key findings: ▸ Figure 1 (left): models are good at using information located at the very beginning or end of its context, but significantly worse at the middle. This isn't unique to GPT architecture - encoder-decoder like Flan-T5 also suffers in the middle ▸ Figure 2 (right): models that are natively longer context do NOT actually use the context better. You can see that the curves of GPT-3.5 (4k vs 16k) almost completely overlap. ▸ Model performance substantially decreases as input contexts grow longer, regardless of their native length. We don't need more tokens. We need models that actually pay attention to them (pun intended).

@TateTheTalisman @DrLoupis 😅😅 savage

@DrLoupis Dude holds rag

Caption this.

@elonmusk It would be insightful to hear some improvements / thoughts Twitter could come up for promoting content for musicians / artists . We are a huge community , and it would be great to have some tools to help us !

Must confess to a deep love of fashion & architecture

Died 🤣🤣🤣🤣🤣🤣🤣

Meme Dealer. Data Dealer.

One Judge decided we are free one judge decided we are not. So we must now return to court, for the second time today and appeal to new judges.

Meta's new app was built entirely using this keyboard:

F*ck threads. We already have Twitter

@TateTheTalisman @Waller7J 😂😂😂

@Waller7J What the fuck is a “soccer”? 🤷🏽‍♂️

Off to watch The Beautiful Game soccer event. 🌴⚽️ American AF, but I know ZERO about the game. Just know my boy is playing. All that matters. 💪🖤💯