> SSA replaces the O(n²) dense attention pass with a learned sparse formulation that scales linearly with context length. SSA's advantage over dense attention grows as context length increases. At 1M tokens, SubQ requires 64.5x fewer compute than dense attention
@blknoiz06 Their post: https://t.co/abrcNf4FTF reads just like DoubleZero's white paper, nonstop AI slop e.g "It's not X it's Y". Both teams are great, but that's an immediate turn off. They could spend some time and actually type their content out.
I've got an agent in a loop optimizing a renderer with the goal to minimize frame times (and tests to measure). It got times down from 88ms to 2ms and allocations down from ~150K to 500. Sounds good, right? Wrong. This is exactly why agent psychosis is a big fucking problem.
As an experiment, I rewrote the Ghostty core render state in Go, with access to identically laid out data structures as Ghostty and the exact same validation tests. I made a purposely naive renderer (simple, correct, but slow). 88ms per frame with 150,000 allocations (horrendous, lol)!
I then kickstarted a Ralph loop to bring the frame times down. I told it it can't modify input data structures or the public API or tests (they're correct), but it can do anything else it wants. It got to work.
It has worked for about 4 hours. I've spent around $350 on this experiment so far. The results?
88ms => 1.5ms
150K allocs => ~500 allocs
Incredible right? Nope.
My hand-written renderer I ported has frame times (same benchmark) of ~20us (0.020ms) and 0 allocations in the update path.
This is the problem with psychosis and lacking systems understanding. If you don't understand the system, you're going to accept that this is an incredible result. If you understand the system, you'll see better solutions immediately and can do roughly 75x better on throughput.
The people who blindly trust agent output are in the former camp. They're sheeple, overdrinking from a fountain of mediocrity.
Standard disclaimer: I use AI all the time. I like AI. The point I'm making is to not blindly accept results. Think. Analyze. Learn.
Since last night, first responders and all necessary services have been working wherever needed. As of now, at least 83 people have been confirmed injured since midnight. Tragically, there are fatalities. My condolences to the families and loved ones.
It was a heavy attack – 90 missiles of various types, many of them ballistic missiles – 36 in total. There were 600 drones. Unfortunately, not all of the ballistic missiles were intercepted – the largest number of hits was in Kyiv. Kyiv was the primary target of this Russian attack.
Putin can’t even pronounce the word “hurrah” clearly anymore – slurs and mumbles – yet he is still vanquishing residential buildings with his missiles. Launched three Russian missiles against a water supply facility. Burned down a market. Damaged dozens of residential buildings. Hit several ordinary schools. Launched his “Oreshnik” against Bila Tserkva. They really are unhinged.
It is important that this does not pass without consequences for Russia. Today, everyone in the world who will not stay silent and chooses to help Ukraine is a defender of life. It is critically important to continue working to secure air defense for Ukraine, especially anti-ballistic capabilities.
We are doing our best to achieve peace and protect people – everything. It is important that Ukraine is not alone. Decisions are needed – from the United States, from Europe and others – to make that old “Oreshnik” in Moscow finally utter the word “peace.”
14 maja 2014, komik Zełenski robi bekę z putina dla swojego programu rozrywkowego, będąc na Placu Czerwonym
Maj 2026, prezydent Ukrainy Zełenski robi dekret o paradzie w Moskwie i zezwala rosji na paradę.
Kino.
nobody will remember:
- your salary
- how “busy you were”
- how many hours you worked
people will remember:
- if your girl has cannons
- if you pay with half stock half cash
- if your microbiome score is top1%
@mageeclegg I think you are just wrong by an order of magnitude. Perhaps the lives of those in the top 0.1% are as great as you make it, but I've never met any.