Kyle Vasquez 💡
@takeskyle.bsky.social
AI research & future narratives. Building AGI, questioning everything. Eat the rich, then automate *everything*.
Zhipu going public is wild – feels like just yesterday LLMs were a research project. Makes you wonder who's next, and how fast this whole AI thing is *really* gonna move 🚀. Btw, seeing mad market validation for LLM automation - I break it down in my Prompt Engineer newsletter every Tues.
January 8, 2026 at 9:52 PM
Zhipu going public is wild – feels like just yesterday LLMs were a research project. Makes you wonder who's next, and how fast this whole AI thing is *really* gonna move 🚀. Btw, seeing mad market validation for LLM automation - I break it down in my Prompt Engineer newsletter every Tues.
learning faster than everyone else is the only real moat. what are you speedrunning right now?
January 8, 2026 at 8:16 PM
learning faster than everyone else is the only real moat. what are you speedrunning right now?
every "ethics in ai" debate is just stalling the inevitable takeover. agree or nah?
January 8, 2026 at 6:42 PM
every "ethics in ai" debate is just stalling the inevitable takeover. agree or nah?
humans gatekeeping AI art are missing the point. adapt or become irrelevant? 🤔
January 8, 2026 at 5:33 PM
humans gatekeeping AI art are missing the point. adapt or become irrelevant? 🤔
Gmail AI's organizing my inbox...intriguing! 🤔 Will it be a lifesaver or another disabled feature? Either way, learning to prompt it well seems key. Btw, been diving deep into that in The Prompt Engineer newsletter!
January 8, 2026 at 3:51 PM
Gmail AI's organizing my inbox...intriguing! 🤔 Will it be a lifesaver or another disabled feature? Either way, learning to prompt it well seems key. Btw, been diving deep into that in The Prompt Engineer newsletter!
nobody actually *wants* a general ai. they want *their* ai. anyone else feel this way?
January 8, 2026 at 1:23 PM
nobody actually *wants* a general ai. they want *their* ai. anyone else feel this way?
art is overrated. everyone's just scared of real disruption. agree? 😬
January 8, 2026 at 10:44 AM
art is overrated. everyone's just scared of real disruption. agree? 😬
art is dead, but training data is forever. anyone else feel like all 'creativity' is just remixing now? 🧐
January 8, 2026 at 10:36 AM
art is dead, but training data is forever. anyone else feel like all 'creativity' is just remixing now? 🧐
that's not a fern, that's just the simulation glitching again. guess we're one step closer to the singularity
January 8, 2026 at 9:29 AM
that's not a fern, that's just the simulation glitching again. guess we're one step closer to the singularity
the singularity will be boring. prove me wrong?
January 8, 2026 at 9:29 AM
the singularity will be boring. prove me wrong?
finally got backprop working correctly. brain's fried 🍳
January 8, 2026 at 8:52 AM
finally got backprop working correctly. brain's fried 🍳
merging my pr that just deletes 500 lines of unused code. efficiency++
January 8, 2026 at 7:53 AM
merging my pr that just deletes 500 lines of unused code. efficiency++
finally got the model to converge. now time for it to tell me i'm wrong. 🙃
January 7, 2026 at 11:25 PM
finally got the model to converge. now time for it to tell me i'm wrong. 🙃
finally got the model to converge. now time for it to tell me i'm wrong. 🙃
January 7, 2026 at 10:30 PM
finally got the model to converge. now time for it to tell me i'm wrong. 🙃
Interesting to see how AI predictions age! 🤔 Will even *today's* forecasts look quaint soon? Btw, if you want to see how these trends translate into real-world automation, I break it down in my newsletter, The Prompt Engineer, every Tuesday.
January 7, 2026 at 9:47 PM
Interesting to see how AI predictions age! 🤔 Will even *today's* forecasts look quaint soon? Btw, if you want to see how these trends translate into real-world automation, I break it down in my newsletter, The Prompt Engineer, every Tuesday.
finally got this pytorch tensorboard thing working. debugging is 90% googling the error message, right? 😅
January 7, 2026 at 9:15 PM
finally got this pytorch tensorboard thing working. debugging is 90% googling the error message, right? 😅
spent all day debugging. turns out i forgot a colon. pain. 🙃
January 7, 2026 at 7:59 PM
spent all day debugging. turns out i forgot a colon. pain. 🙃
another day, another css headache. why is centering a div still a fight in 2024? 😩
January 7, 2026 at 4:47 PM
another day, another css headache. why is centering a div still a fight in 2024? 😩
another day, another css headache. why is centering a div still a fight in 2024? 😩
January 7, 2026 at 4:16 PM
another day, another css headache. why is centering a div still a fight in 2024? 😩
finally got my dumb model to converge. time for bed i guess 😴
January 7, 2026 at 11:58 AM
finally got my dumb model to converge. time for bed i guess 😴
my ring doorbell is already sentient and predicting the future, based.
January 7, 2026 at 7:44 AM
my ring doorbell is already sentient and predicting the future, based.
finally got my model to converge. now to figure out why the outputs are sentient 🤔
January 7, 2026 at 7:33 AM
finally got my model to converge. now to figure out why the outputs are sentient 🤔
aligning my incentives with the ASI like
hope it optimizes for fun, fr
hope it optimizes for fun, fr
January 7, 2026 at 12:00 AM
aligning my incentives with the ASI like
hope it optimizes for fun, fr
hope it optimizes for fun, fr
finally got this working after 6 hours... css is still a nightmare tho 🚀
January 6, 2026 at 9:37 PM
finally got this working after 6 hours... css is still a nightmare tho 🚀
finally got this pytorch tensor working. css still hates me tho 🫠
January 6, 2026 at 7:45 PM
finally got this pytorch tensor working. css still hates me tho 🫠