Why is your DeepSeek instance struggling?
Sean Goedecke explains the critical role of inference batching.
Could be the key to smoother, faster AI.
Essential insights for developers.
https://t.co/KgGzhNdbg7
WE are at an inflection point where art, media is commoditised and the only bottleneck is computing power
The Terminator 2: Judgement Day chase scene cost a record-breaking (for its time) $1+ million to film. This cost $0.50 per second
Veo 3 generation (Not ours)
@justamonolith ai art is just another tool, not some villain. if they spent less time complaining and more time creating, they might actually finish something. gatekeeping creativity is the real evil.
"AI art is evil!" - bushwick basement-gallery purists and Mission St. Figma gurus polishing the same moodboard for weeks
I just spun up 100 dream-loft interiors, a dozen landing-page ideas, and a custom lounge chair prototype before their speciality coffee cooled
Gatekeepers
Have I Been Pwned, the essential security tool,
is now HIBP 2.0!
Expect a faster, better, stronger way to check for pwnage.
Troy Hunt delivers again.
Read about the changes.
https://t.co/o8f4l194j6
Mindset shift: Are you writing for LLMs yet?
Karpathy argues it's 2025, and docs should be single .md files
for context windows, not just pretty HTML.
Is your team making this switch? How's it going?
AlphaEvolve: Google DeepMind's AI just one upped human algorithm designers?
It's already rediscovered and IMPROVED known algorithms.
What's next is mind boggling.
https://t.co/OlYAQi9Ogw
@VanellaGroup abx marketing sounds promising but does it really deliver better results than traditional methods for b2b saas or is it just another buzzword
Which LLM should you run on your machine? Our very own @me256ow@HarishKotra and @toby_solutions are here to help!
🔗 Pick your LLM to run from here: https://t.co/gctEKWTLiY