Nicolas Pratt

Weirded out by homes with no pets. Like where is your creature

every millennial whose love of basketball is due in part to Deron Williams and Dee Brown is smiling ear to ear right now. Long time coming.

ARE YOU WITH US? For the first time since 2005, we're heading to the FINAL FOUR.

Just good old B1G smack talk

Illinois vs Iowa

This is the most Trump quote of all time. It sums him up perfectly. It also explains his cabinet.

“What did he whisper to you?” “I can’t tell you” “Come on, I won’t tell anyone” “No I can’t. Secret, secret” David Mirkovic is box office

I think the worst part of the last 10 years was that it was so obviously avoidable to anyone that reads above a 4th grade level.

Nothing says “Eastern Europe” quite like Champaign-Urbana, Illinois

@markoilico Would love this!

If you're now designing or redesigning a website, this will help you a lot. I recently curated the best hero sections, footers, social proof and other website parts because I got tired of having 15+ tabs open (even with Mobbin). Giving it away 100% free. Comment on this post, and I'll send a Figma link to your inbox!

Every time you message an AI chatbot, the model stores your entire conversation in temporary memory called a KV cache (a cheat sheet so it doesn’t re-read everything from scratch). On a large model like Llama 70B running a long conversation, that cache alone eats 40GB of GPU space, often more than the AI model itself. That’s half a $30,000 GPU chip consumed by one user’s memory. Google just published TurboQuant, a compression algorithm that shrinks this cache by 6x, down to just 3 bits per value, with zero accuracy loss across every benchmark tested. No retraining. No fine-tuning. Drop-in replacement. AI inference (running models for actual users, not training them) now makes up 55% of all AI compute spending. Hyperscalers are pouring nearly $700 billion into AI infrastructure in 2026. The KV cache is the single biggest memory bottleneck in that stack. When GPU cache memory fills up, the system can’t take more users. 6x compression means the same hardware handles roughly 6x more simultaneous conversations, or 6x longer context windows, or some mix of both. At cloud rates of $2-3/hour per H100 GPU, that’s the difference between profitable and unprofitable AI deployment. TurboQuant randomly rotates data to simplify its structure, applies a compressor, then adds a 1-bit error correction step to catch errors before they compound. On H100 GPUs it delivers up to 8x speedup over uncompressed computation. Google tested it across five long-context benchmarks on Llama, Gemma, and Mistral models. Perfect scores on needle-in-a-haystack (finding one specific fact buried in massive text). Being presented at ICLR 2026. It also outperforms existing methods for vector search, the technology that powers how search engines find similar results across billions of entries. Google runs billions of these searches daily. Three bits. Zero loss. 6x compression on the biggest memory bottleneck in a $700 billion infrastructure buildout.

Just to be clear, we already have free data centers. They are called libraries.

RIP Sora, you gave us the greatest ai video of all time

RT @rjonesy: The same clarity @Flighty bring to flights... now for airports. Plus a free version at flighty.com/aiports for monitoring…

it's beautiful that I can traverse through 25 years of UI design history by clicking 3 times on Windows 11.

Cats

You weren’t spending 5 hours at the airport under Biden….matter of fact….the airlines were required to pay you money if they messed up…..just food for thought.

The best time to plant trees is 20 years ago. The second best time to plant trees is today

there are two truly american art forms: jazz and eating in your car

The only one. No comparisons.