Alex Krieger (@kregasaurusrex) header
Alex Krieger (@kregasaurusrex) profile picture

Software engineering/ML in Python, grinding leetcode, beating financial algos, and spending too much time in cyberspace. Speedrun the tech tree! #dogecoin

Cleveland, OH
Joined April 2012
Photo by Alex Krieger (@kregasaurusrex)Photo by Alex Krieger (@kregasaurusrex)Photo by Alex Krieger (@kregasaurusrex)Photo by Alex Krieger (@kregasaurusrex)Photo by Alex Krieger (@kregasaurusrex)Photo by Alex Krieger (@kregasaurusrex)Photo by Alex Krieger (@kregasaurusrex)Photo by Alex Krieger (@kregasaurusrex)Photo by Alex Krieger (@kregasaurusrex)Photo by Alex Krieger (@kregasaurusrex)
Alex Krieger retweeted
I just built the world's fastest SAT-solver (and it's not even close), thanks to Neon. Fifteen years ago I spent five years (thousands of hours) working on an advanced Boolean circuit preimage solver, aka a SAT-solver. Building an algorithm to efficiently solve a large subset of the NP-complete problem space in polynomial time is extremely difficult. My solver worked, but it was very slow compared to the state of the art. Today I gave Neon all my 15-year-old solver code. It analyzed all of the code, took all my best unique ideas and suggested how to improve them using other algorithms, and then it built a highly efficient hybrid solver out of all of the ideas. The result was FAST: MUCH faster than what I had built. This whole process took LESS THAN FIVE MINUTES. I spent FIVE YEARS in the muddy trenches of number theory and cryptography, and I only got 5% of of the way to where Neon got within 5 minutes. I gave Neon a handful of ideas as to how to improve its initial algorithm, and two minutes later the solver was 50x faster. I gave it more ideas: 10x faster again. More ideas: 100x faster again. This new solver, Frobenius, is an alien speed demon. I'm only 20 prompts deep, and Frobenius is already the fastest SAT solver in the world, by a large margin. Frobenius is an average of thousands of times faster than state of the art solvers across all problems, but the gap grows even wider as the problem size increases.🤯 This is how all future progress will be made: through human ⇄ AI symbiotic feedback cycles. We are now cyborgs: augmented human-AI-hybrids. And a human plus an AI will always accomplish more than either individually. I can't help but think how much more impact I could have had during my career if I had these LLM tools 20+ years ago. P.S. 1: I'm not ready to release Frobenius yet, because the big SAT solver competition is in April 😁) P.S. 2: Anthropic is in trouble. Neon is 10x more powerful and 100x faster for hard tasks like this than Opus 5.5. I just canceled three out of my four Claude Ultra plans.
92
148
31
1,477
76,230
Alex Krieger retweeted
You're kidding
167
338
60
14,785
1,511,720
Alex Krieger retweeted
This is real-time AI video generation (no gaming engine!) using just half of one AI4 TRIP accelerator. Let that sink in …
The crazier thing is that the right side uses less compute than what the left used to take. Extreme compute efficiency optimization.
1,143
1,467
293
16,388
1,774,579
Alex Krieger retweeted
Update: They made up!!
A Georgia Tech couple appears to have a breakup live on air and the commentators analyze and break it down #StingEm #GoDeacs
442
3,681
543
69,011
3,271,887
Alex Krieger retweeted
A Georgia Tech couple appears to have a breakup live on air and the commentators analyze and break it down #StingEm #GoDeacs
765
6,629
1,279
137,843
16,532,704
Levent Alpoge writing about claude:
I abandon myself to the crosswalk signal. And I am rewarded with the appearance of a friend, the little man made of light
3
122
1,967
32,614
Alex Krieger retweeted
This user found his apartment in GTA VI [📹 No-Lock216]
67
482
31
4,857
231,787
Alex Krieger retweeted
How I make small talk at the formal function so I can avoid any questions about myself
64
1,186
89
11,323
718,548
6
57
5
1,471
30,550
Alex Krieger retweeted
Replying to @rawsalerts
9
38
3
600
47,646
🚨 #BREAKING: Mathematicians conflicted over the emerging field of “cuckulus” whereby mathematicians prompt AI to solve unsolved problems while they just sit back and watch.
390
3,245
599
43,985
2,440,913
Alex Krieger retweeted
🚨 A TEXAS SCHOOL DISTRICT HAD A PLAN TO SAVE MONEY BY HIRING FOREIGN TEACHERS. The plan wasn’t to pay them less. It was labor-cost arbitrage. They’d avoid payroll taxes, retirement contributions and health insurance expenses the district expected to pay for American teachers. Don’t take my word for it. Granbury ISD put the math in writing: $10,000 saved on FICA, health insurance and TRS retirement. $9,000 paid to an international recruiting middleman. $1,000 supposedly saved for taxpayers. Sounds like a great deal, right? Except the math didn’t hold up. Nine months later, in March 2026, Granbury returned to its school board with an embarrassing correction. Oops. We still have to pay TRS. The district acknowledged that it remained responsible for retirement contributions. It also confirmed that Granbury pays for IAG health insurance when teachers elect that coverage. Two major components of the original savings pitch were now in question. So where was the new financial analysis? The public correction contained no revised savings calculation. Instead, officials asked the board to continue approving the recruiter’s invoices. And here’s where it gets even worse. They STILL hadn’t addressed the Social Security problem. Granbury’s own payroll guidance says most district employees are already exempt from Social Security taxes. You cannot save 6.2% in employer Social Security taxes by importing a J-1 teacher if you weren’t paying that 6.2% on the comparable local teacher in the first place. So I reconstructed their original savings pitch using publicly available rates. Here’s how the supposed savings collapse: $10,326 - A reconstruction of all four originally claimed savings components. $4,590 - After removing TRS and unsupported health insurance savings. $870 - After removing Social Security savings the district generally wasn’t paying anyway. That’s it - Just $870 in possible employer Medicare savings at the earlier $60,000 salary, assuming the J-1 teacher qualifies for the exemption. NOW LOOK AT WHAT THEY AGREED TO PAY THE MIDDLEMAN. IAG’s contract charges 20% of a teacher’s annual salary during the FICA-exempt period. At Granbury’s $61,000 starting salary for 2025–26, that’s $12,200 per teacher, per year, before applicable credits. Using the district’s published health insurance contribution and a modeled TRS credit, I compared the first-year cost of hiring a local teacher with a J-1 hire through IAG. LOCAL TEACHER: $67,640.50 J-1 TEACHER VIA IAG: $77,736 THAT’S $10,095.50 MORE PER YEAR. Not the $1,000 in savings Granbury originally advertised. This is just my model - the best I can do with what I have, it’s not verified district invoices. But it raises an obvious question: Where is Granbury’s actual math showing this arrangement saves taxpayers anything? But wait there’s more. IAG also advertises fees to recruited teachers, meaning the intermediary potentially collects from BOTH the taxpayer-funded school district and the teacher. Think about what happened here. A public school district justified international recruitment partly by claiming it could avoid employer taxes and benefits. Nine months later, it acknowledged a major mistake and now I’d like some answers. @granburyisd: Will you release the actual IAG invoices, insurance premiums, credits and updated per-teacher cost comparison? How much has this arrangement actually cost taxpayers? @GinaHinojosaTX: Will you examine whether Texas school districts are spending MORE to recruit internationally than they would to hire comparable teachers locally? Because if taxpayers are paying extra to recruit teachers overseas while being told the arrangement saves money, that’s a story every Texas parent deserves to know about.
Texan Parents, Take a couple minutes to read through these bios and then ask yourself: Is it in our children’s best interests to import foreign teachers on visa? A nation which cannot educate their own youth, is not a nation at all. @GinaHinojosaTX we can do better than this.
123
901
63
2,442
96,561
Alex Krieger retweeted
I don't understand how at Rockstar: - they have so many draw calls (individual parts) & materials in view - they have so many hair simulations running at once - they have so many volumetric lights casting soft bokeh shadows - they have so many rigidbodies (every vehicle, creature, prop, door, rock, plant is a physical object). - they simulate so many unique NPCs (pathfinding, AI, voices, animation, hair simulation, active ragdoll, relationships, possessions, etc...) - they have global illumination with no noise and no barely any temporal artefacts - they have full raytraced reflections - they stream all of this data seamlessly - they render this all at native 1440p, no dynamic resolution It's a monumental achievement, a lesson for any gamedev, like Red Dead Redemption 2 was & still is. That's what the hardware is capable of. That's the bar.
I’ve made video games for 19 years and GTA VI appears to me as a Bugatti Veyron would to a nervous Victorian child.
189
745
67
13,256
557,248
Alex Krieger retweeted
Trump endorsing Stephen Miller for President over JD Vance and Marco Rubio needed a soundtrack. Wait for it, and watch Vance and Rubio both realize they sold their souls for nothing.
1,549
6,451
876
44,281
3,105,253
Alex Krieger retweeted
Aschenbrenner > burns things to ashes in german. Don’t ever doubt nominative determinism
RUMOR: Leopold Aushenbrenner was liquidated today for the third time in a row.
47
153
10
3,506
96,158
Alex Krieger retweeted
71
472
44
4,529
184,975
Alex Krieger retweeted
「なんで青色?」 って思ったけど完成見て腰抜けた...
33
1,021
58
10,317
1,132,021
Alex Krieger retweeted
Y'all remember when CJ used to receive calls for being too fat by legit BIG Smoke lmao
This is how fat Jason and Lucia can get in GTA 6
62
1,574
51
25,767
531,017
Replying to @yoracrab
2
223
1
8,477
152,122
Alex Krieger retweeted
Props to Rockstar, GTA 6 shipped before the Riemann Hypothesis was proven!
224
305
42
7,543
3,993,094