←
Alexander PCs4,057 posts
Alexander PCs avatar

Alexander PCs

@AlexanderPCS

Expertly Crafted Custom PCs. AI and data science for some of your favorite (and least favorite) teams and players. NVIDIA Partnered. 🇺🇸 Veteran Owned.

◷ Joined April 2020▧ 1,356 Photos and videos
Alexander PCs avatar
Alexander PCs@AlexanderPCS·16h

This has been on our minds all afternoon, and we wanted to take some time and think about an appropriate response to it… We want to begin by saying that NVIDIA has been a wonderful partner to our company from literally the very first moment we started working together 6+ years ago. We have enormous respect for what Jensen Huang built, the people who work there, and the extraordinary contributions NVIDIA has made to American technological leadership, and to our business specifically. Jensen Huang is a singular entrepreneur in our space, and we’ve been huge fans of him and his work for many years. Jensen’s personal mistakes are ultimately matters for him and his family. We have no interest in passing judgment on his marriage or his personal life. None of us is perfect. That being said, the allegations surrounding Megaspeed, the potential diversion of restricted AI chips into China, and the possibility that individuals close to NVIDIA’s leadership were involved are matters of an entirely different magnitude. These are not simply questions of personal judgment. They are questions of corporate integrity and American national security. If Jensen was unaware of any wrongdoing, a thorough independent investigation should establish that, and his name should be cleared. We would really, really, really like for that to happen. But if he knew, facilitated, or deliberately ignored the circumvention of American export controls, then he should never return to NVIDIA’s leadership. The same standard must apply to every executive who may have been involved. NVIDIA is bigger than any one individual, including its founder. Its employees, customers, partners, shareholders, and our country deserve a complete accounting of what happened. We sincerely hope that Jensen’s legacy survives this. But protecting his legacy cannot take precedence over discovering the truth. We remain proud to be an NVIDIA partner, and look forward to continuing to work with the brilliant people and products that we love and appreciate. Our partnership is with a company and its people, not any single executive.

Over the past week, I’ve been overwhelmed by the love and support of friends, industry colleagues, and people I've never met. Thank you to everyone who has shown kindness to my family and me. It means the world to me. I came to the United States 53 years ago, and becoming an American citizen is one of the greatest gifts and proudest moments of my life. NVIDIA has been my life’s work. Building it into the company it is today alongside my colleagues has been a privilege and an honor. Nowhere but in America could a company like NVIDIA be possible. I love our country, and I love NVIDIA. I did not live up to the standards expected of everyone at NVIDIA, and I have made serious mistakes in my personal life. I created problems for the Board that I deeply regret. I own my mistakes completely, and I am committed to cooperating with the Board’s independent investigation. Since my departure, I’ve seen an outpouring of calls from NVIDIA’s employees, investors, business partners, and customers for me to return as CEO. Many continue to reach out to me directly as the company navigates complex technology, market, competitive, and geopolitical dynamics. NVIDIA is a once-in-a-generation company. If the Board of Directors decides that my return is in the best interest of our shareholders, our employees, the company, and America, I would be honored to answer these calls and return as CEO. Thank you again to everyone who has shared their good wishes. I am profoundly humbled by your support. To my family: I love you, and I will spend the rest of my life being worthy of yours. Jensen

Alexander PCs avatar
Alexander PCs@AlexanderPCS·Nov 17

#15 - Andrew Miller - At his peak, an absolutely terrifying relief pitcher. Revolutionized how managers thought about deploying elite bullpen arms. Unfortunately, being unhittable for a few Octobers isn't a complete Hall résumé. #16 - Omar Vizquel - Eleven Gold Gloves and nearly 3,000 hits. His supporters see a defensive wizard. His detractors see a 45-WAR player who hung around forever. We’re with the latter. Tier 5 - “Hall of Very Good. General Admission. No Reentry.” #17 - Wade Davis - His 2014–2015 run was absurd. A 0.94 ERA in 2015. But the Hall of Fame is not the Hall of Two Incredible Seasons. #18 - Jake Arrieta - His 2015 season was one of the greatest pitching performances of the century. His overall career? Not remotely close. Cooperstown doesn't offer single-season memberships. #19 - Kyle Seager - 36.9 WAR, 242 home runs, excellent defense at third. A very good player who had the misfortune of playing most of his career on Mariners teams that nobody outside Seattle watched. #20 - Brett Gardner - Nearly 45 WAR! Genuinely underrated. A valuable leadoff hitter and defender. But being better than people remember doesn't make you immortal. #21 - Ryan Zimmerman - The face of the Nationals for years. Nearly 1,850 hits, 284 home runs, and a World Series ring. Deserves a statue in Washington, maybe. But definitely not a plaque in Cooperstown. #22 - Jay Bruce - 319 home runs and three All-Star selections. Good power hitter. Absolutely no Hall of Fame case. Tier 6 - “Thank you for your service. Please enjoy the museum as a paying customer.” #23 - Asdrúbal Cabrera - Two All-Star appearances, a solid 15-year career, and one of the coolest unassisted triple plays you'll ever see. His Hall of Fame case begins and ends with someone accidentally opening the wrong spreadsheet. #24 - Starlin Castro - Nearly 1,800 hits, four All-Star selections, and a long career as a perfectly serviceable infielder. Once led the National League in hits. #25 - Ervin Santana - 151 wins, over 2,000 strikeouts, and a no-hitter. An admirable career. Tier 7 - “Anyone who votes for this guy should have their ballot confiscated.” #26 - Pablo Sandoval - Three rings, a World Series MVP, and one of the most memorable three-homer World Series games ever. Also ~20 career WAR. The Panda belongs in Giants folklore, but not the Hall of Fame. #27 - Joakim Soria - A legitimately good closer with 229 saves and two All-Star selections. But if you're voting for Joakim Soria for the Hall of Fame, we’re going to need you to explain what you think the Hall of Fame actually is. So that’s the list. But if WE had a ballot, who would we actually vote for? This year, our ballot would have two names on it. Alex Rodriguez, and Chase Utley. Let us explain… Alex Rodriguez - he’s so much better than anyone else on this ballot it’s absolutely insane to not vote for him. Also, they’re not going to vote for him. Buster Posey - Undoubted Hall of Famer. But not, for us, a first ballot career. So we’d vote for him, just not this year. Doesn’t have the longevity for a first ballot nod from us. He’s getting in first ballot anyways. Chase Utley - we get it, the 1885 hits are a legitimate objection. We’re ultimately giving Utley the nod for his insane 7-year peak. Utley’s 7-yr peak bWAR is 49.3 which is significantly above the HoF position average. He was elite defensively, and put up over 250 HRs which is great power for a 2B. Career JAWS of 56.9 is right at HoF standard. Andy Pettitte - career ERA of 3.85. Career postseason ERA of 3.81. Gets a lot of love for having the most postseason wins. But he was the same pitcher in the postseason as he was in the regular season. And his great teams carried him in both. Not saying he wasn’t good. Just saying he wasn’t Hall of Fame good. Lester & Hamels - Not, in our opinion, Hall of Famers. There’s no world where Curt Schilling doesn’t get it, and Lester and/or Hamels does. At least there shouldn’t be. 🧵2/2

Alexander PCs avatar
Alexander PCs@AlexanderPCS·Nov 17

The 2027 Hall of Fame Ballot: A Completely Unofficial Voter's Guide 27 players. Seven tiers. From “put him in Cooperstown!” to “wut?” We’re judging careers, not popularity. Peak performance, longevity, positional value and historical significance all count. And yes, we’re willing to elect PED users. Tier 1 - “Build the plaque. Stop wasting everyone's time.” #1 - Alex Rodriguez. 117.5 WAR, 696 homers, 3,115 hits, three MVPs. If you can't tell the story of baseball without a guy, he belongs in the Hall. Put an asterisk in the corner or the plaque if you want. Whatever. This man is one of the 20 best ball players who ever lived and it’s insane he’s not in. (Also, he’s not going to get in.) #2 - Buster Posey. Three rings, an MVP, a batting title, and one of the best catching peaks in modern baseball. Shorter career, but catchers shouldn't be judged like first basemen. He’s going to be a first ballot Hall of Famer. Tier 2 - “He’s getting in, but we’re going to argue about it first.” # 3 - Chase Utley - Roughly 65 WAR at second base. Elite peak, exceptional baserunning, outstanding defense. If you're voting against him because he didn't get 2,000 hits, please return your calculator to 1987. Utley’s day will come, but it may not be this year. It’ll be close. #4 - Félix Hernández - King Félix was arguably the best pitcher in baseball at his peak. A Cy Young, two ERA titles, six All-Star selections. His decline was ugly, but the peak was spectacular. This will be his 3rd year on the ballot, and he’s on track to be close this year if his trajectory holds. #5 - Cole Hamels - Approximately 59 WAR, 2,560 strikeouts, a 123 ERA+, and a World Series MVP. One of the most consistently excellent pitchers of his generation. A stronger case than casual fans realize. This is his first year on the ballot, so he’s not going in this year. But he eventually gets there. Probably. #6 - Jon Lester - Three championships, 200 wins, a 2.51 postseason ERA. A great big-game pitcher. His career WAR hurts, but the narrative is almost tailor-made for Hall voters. We expect a long, but ultimately successful campaign. Tier 3 - The Hall of Very Good, Executive Level #7 - Andy Pettitte - 256 wins, five rings, 19 postseason victories. A tremendous career, but the Yankee mythology is doing some heavy lifting here. Career ERA of almost 4. No Cy Young. No ERA title. Cooperstown is about recognizing great individual players. If Pettitte gets in, it would be because he had great teammates. #8 - David Wright - Had a Hall of Fame trajectory until his back betrayed him. Approximately 49 WAR and a tremendous peak. Inexplicably had an MVP stolen from him by Jimmy Rollins. The Hall of Very Good has his plaque ready. #9 - Dustin Pedroia - MVP, Rookie of the Year, four Gold Gloves, three rings. His peak absolutely belongs in the conversation. Unfortunately, his knees filed for retirement before the rest of him. #10 - Bobby Abreu - Nearly 60 WAR, 400 stolen bases, 288 home runs and a career .395 OBP. A sabermetric darling whose biggest crime was being excellent without being particularly famous. #11 - Mark Buehrle - 214 wins, 3,283 innings, a perfect game and remarkable durability. The last man in baseball who apparently believed starting pitchers were supposed to pitch. #12 - Jimmy Rollins - A stolen MVP, four Gold Gloves, 2,455 hits. An outstanding shortstop. Also a career 95 OPS+. The MVP trophy doesn't magically turn an okay hitter into a great one. No soup for you. Tier 4 - “I won't vote for you, but I won't mock anyone who does.” #13 - Francisco Rodríguez - 437 saves, including a record 62 in 2008. If you believe closers belong in the Hall, K-Rod has a real argument. We’re not just not convinced. #14 - Torii Hunter - Nine Gold Gloves, 353 home runs, 2,452 hits. A terrific player and an even better highlight reel. His defense made SportsCenter, but his total value falls short. 🧵 1/2

MLB@MLBNov 16

14 first-timers and 13 returnees are on the BBWAA ballot for the Hall of Fame Class of 2027! Who do you think will make it to Cooperstown?

Post attachment
Alexander PCs avatar
Alexander PCs@AlexanderPCS·Nov 16

Senator, Congress controls the purse strings. You’ve served in Congress for 45 years. If America’s aviation infrastructure has been neglected for decades, you don’t get to pretend you’re innocent. It wasn’t an air traffic controller staffing shortage. A router installation flooded the FAA’s aging communications network with connection attempts, disrupting radar and communications across the country. Hiring another thousand controllers or technicians wouldn’t have stopped that network failure. The Trump administration doesn’t get a free pass either. We’re spending a *TRILLION* dollars annually on “defense” yet we’re being told we need additional taxpayer funding to keep the nation’s civilian aviation infrastructure functioning? Before demanding more money, explain why modernizing critical infrastructure isn’t already a higher priority. Congress controls appropriations. The administration controls execution and proposes spending priorities. Both of you have responsibilities here. Stop pointing fingers. Fix. The. Fucking. Infrastructure. Or go away to the nursing home so we can get someone who will.

We just saw the third nationwide FAA ground stop in US history, and instead of assuaging the flying public’s concerns, the Trump admin is pointing fingers. Thousands of Americans slept on airport floors because the Trump admin has been asleep at the wheel and deflects all blame. @SenateDems will use every tool we have to hold them accountable and ensure this does not happen again.

Alexander PCs avatar
Alexander PCs@AlexanderPCS·Nov 16

The data are really strong. Let’s take a look… Baseball has the highest natural variance of any major sport. In baseball, when the best team plays the worst team, the worst team wins about 32% of the time. So upsets in baseball aren’t surprising. Lesser teams beat better teams all the time in baseball. This is opposed to football and basketball where the best team beats the worst team about 96-97% of the time. So baseball *looks* balanced because it has the highest natural competitive balance of any sport. Let’s look at the number of champions from 2006 through 2025 - the last 20 seasons for each sport. Baseball has 12 different winners. Football and hockey have 13 each. Basketball has 11. So it looks like a cap doesn’t matter if you look only at the winning team. All the sports seem to have about the same number of winners regardless of if they have a cap or not. But looking ONLY at the number of different winners doesn’t tell the story of competitive balance across the whole league. You have to look at results vs payroll spending and results vs market size across the entire league in order to have a more complete picture of that. If you look at that data, the top 5 markets in MLB have won the most games over the last 20 years. The largest-market teams outperformed the smallest-market teams by 4.3 percentage points in winning percentage. Compare that to the three capped leagues where in all three the smallest-market teams actually outperformed the largest-market teams over the same period. That doesn't prove salary caps caused the difference. But it shows that the relationship between market size and winning is fundamentally different in baseball. The payroll data is even more clear. There is a completely consistent relationship between a team’s payroll spend and the number of wins. Teams with bigger payrolls, win more games over the long run. Teams that spend less, win less. Sounds intuitive because it is, but the data prove it’s true. You can buy wins in baseball. And the biggest teams do. And in a sport with the most natural variance, a couple wins is often the difference between playoffs and no playoffs - and that comes with a lot of revenue. The salary cap argument isn't that every team should win the same number of games. It's that every franchise should have a realistic opportunity to build a winner, regardless of its market size. Baseball already has more natural competitive balance than any other major American sport. Imagine what it could look like if its economic system encouraged that balance instead of working against it.

Post attachmentPost attachment

show us? because baseball has the most unique champions out of any league since 2000. with no salary cap. So show us?

Alexander PCs avatar
Alexander PCs@AlexanderPCS·Nov 16

But this is also an issue where the union can get meaningful concessions from the owners in exchange for agreeing to it. It’s happening anyways. Doesn’t matter if the union wants it or not. They can talk whatever shit they want, they’re getting a cap. So they need to get something back. Preferably several things… #2 - A salary floor. The cap is inevitable, so the union needs to make sure that cap comes with a floor. The union - very correctly - says that some owners are just greedy billionaires who want to sweat the assets and not invest in a competitive team behind the bare minimum to avoid full on fan revolt (looking at you, Castellini). A salary floor forces owners to spend on a competitive roster. This alone isn’t enough, but more on that later… #3 - Revenue sharing. The unions needs to make sure the cap and floor rise with league revenues. There needs to be a 50/50 split between the owners and the players. It needs to be “player compensation” only and not “baseball operations” that counts for the players’ revenue share. Big W for MLBPA. #4 - Roughly double the league minimum to $1.5 million/yr. The union can, correctly, claim this as a win for its members. W for MLBPA. #5 - Roster expansion. Player rest and recovery is a big priority. Instead of cutting games and giving up that revenue, preserve 162 games and give individual players more rest. New roster goes from 26 active players to 30. More union members. Huge W for MLBPA. #6 - Smarter travel. Attack player fatigue with optimized travel and roster expansion simultaneously. These is a meaningful quality of life win for the Union that it can rightly feel good about securing for its members. Beyond salary cap and floor, revenue sharing, and player quality of life changes, there are a number of ideas we have that we believe would enhance the overall baseball experience for the FANS. The union should work these items into the CBA as well… - Get rid of interleague play during the regular season. - Get rid of divisions. It’s the American League, and the National League. - top 4 teams in each league make the playoffs. Playoffs are all best-of-7 series. - Pool all MLB broadcast deals. No blackouts. Teams split 50% of the pool 1 share per team. At a $4 Billion pool, that’s $66.67 million per team equal base payout. The second 50% gets paid out based on performance. The team that wins the most games in each league gets 15 shares. Second place gets 14 shares, on down to the bottom team’s 1 share. At $4 billion that means the team with the best regular season record in each league gets $125 million and the last place team gets $8.33 million. So adding back in the equal shares portion, the top team in each league takes in $191.67 million and the bottom team takes in $75 million. This means WINNING is the only way to have the most financial success. Every single regular season game is now meaningful. It kills tanking and makes winning the best business model. It gives EVERY team a reason to be active in free agency. It also resolves a major problem that the union has regarding mid-tier veterans. Under the current CBA, top free agents get all the free agency money across a few big contracts leaving the mid tier guys with nothing as the rest of the focus shifts to younger players on pre-arbitration contracts and/or under team control. - Remove the draft reward for finishing last. We’ve already hit the owners in the pocketbook by tying half the broadcast revenue to performance. Now we make “rebuilding” a bad business model by giving every non-playoff team equal draft lottery odds. We lottery the entire first and second rounds between picks 1-22 and then lottery 23-30 between the playoff teams. Losing should never automatically improve your draft position, your revenue distribution, or your finances. 🧵 2/2

Alexander PCs avatar
Alexander PCs@AlexanderPCS·Nov 16

Weighing in on @MLB and @MLBPA - seeing so many people get this all wrong. “Players are making tons of money. Guys are making $30, $40 million a year…” Yeah, like 6 guys in all of MLB are making $40 million a year. There’s an average of less than one guy per team making $30 million a year. There’s a huge concentration of wealth at the top of the MLBPA membership. The top 100 highest paid players account for about 49% of all player salaries. About 45% of players made less than $1 million in 2026. The median salary for 2026 was $1.4 million - so more than half the league is making less than $1.5 million a year. And those are opening day roster figures, so when you add in the hundreds of minor league call-ups over the course of the season where those players get prorated shares of the $780k/yr league minimum, you end up with most of the league making less than $1 million a year. This massive earnings disparity among its membership is a problem for MLBPA. And at the negotiating table, it’s a fatal one. Why? Because of how the voting works. While every member of the union gets one vote at the team level - whether they make $40+ million or $780k - only 38 actual votes are cast when it comes to approving a CBA or not. There are 30 team player representatives, one elected player rep from each team, that cast a single vote reflecting the consensus of their clubhouse. Then there are 8 executive subcommittee members - the 8 player leaders elected to the MLBPA subcommittee - who each cast an individual vote. This structure and the massive wage disparity among members mean the union can never win a prolonged work stoppage. The guys with big money can weather a lockout a lot easier than the pre-arbitration league minimum guys and minor leaguers who make up the majority of the union’s voters. (Scott Boras can say whatever he wants… he doesn’t speak for the MLB rank and file, and the MLB rank and file are not Boras clients.) Owners will use this as a wedge. They know they can pair significantly increased minimum salaries with a cap & floor and after a few months of lockout, the lower earners will be ready to surrender. If you’re on $780k/yr but you haven’t had a check in 4-5 months and you’re looking at a deal that will double your pay and get you to arbitration and/or free agency a year sooner… lotta guys are going to vote to take that deal. In fact, more and more guys with every round of negotiations. The MLBPA acts like it doesn’t already know this, but it does. This is a fight they literally will never win. And both sides know it. In fact, this is how the last CBA was passed - when the rank and file clubhouses effectively overruled union leadership and 26 of 30 teams voted, against all 8 executive subcommittee members, to approve the deal and it passed 26-12. So with owners eventually “winning” any prolonged standoff anyways, the union needs to focus on a set of objectives that are achievable. Just because “greedy billionaire” victory is assured (and it IS assured) doesn’t mean the two sides don’t have mutual interests. They have a LOT of mutual interests. And a lot of these things can get done, and many can be genuine wins for the union. #1 - A salary cap is inevitable. It works in other sports and has for years. It helps competitive balance - yes, we know the Brewers and Rays aren’t major market cubs blah blah blah - but the data across other sports prove it’s better over time. And, critically, many fans want it. That’s never been true in previous CBA environs for baseball, but it’s absolutely true now. Fans are sick and tired of the dodgers and their $3 billion payroll with deferred comp paid through the year 2079. They’re sick and tired of the top 4 teams spending more than the bottom 22 teams on free agency - which is what’s happened over the 2022-2026 CBA period. MLBPA needs to understand that a huge portion of the fan base is NOT on their side with regard to this issue of a salary cap. 🧵 1/2

Alexander PCs avatar
Alexander PCs@AlexanderPCS·Nov 15

Hello, police. We’d like to report a murder.

Oh, sweetheart. You’re lecturing the Gold Star wife of a combat infantryman on the distinction between combat and noncombat service. Nobody suggested their experiences were the same. In fact, military compensation already accounts for those distinctions through hostile fire pay, imminent danger pay and hazardous duty pay. What you seem incapable of grasping is that the benefits earned through military service are not contingent upon whether someone was shot at. I’m perfectly capable of honoring my husband’s sacrifice without diminishing someone else’s service. Perhaps you should ask yourself why that’s so difficult for you.

Alexander PCs avatar
Alexander PCs@AlexanderPCS·Nov 15

Wow… LA passed on a chance to make this guy the mayor. Fumbled bag. He’d have been awesome.

We need to start a doomsday panic scorecard. Every time someone declares a civil war, or climate apocalypse, or nuclear armageddon & WW3 for the 9,000th time, it never comes to pass, but we just move on, so 5 months later they do it again, and people fall for it, again. The apocalypse isn't happening in your lifetime, folks. The real shit is some overpaid bureaucrat in your municipal water district is gonna get lazy with their emails, and your reservoir is gonna be empty when some other overpaid state park bureaucrat forgets to monitor a smoldering plant for an entire week, then your entire town burns up in a wildfire, your grandma burns alive and then 400 more people die from accelerated cancers from the toxic smoke. Buying into these theatrical fantasies of global catastrophe only serves our corrupt government, distracting you from the true menace in your life. It's so exhausting. I used to indulge in these catastrophic conspiracy theories, but I had to learn the hard way that it's just plain old government incompetence that's the biggest threat to your daily life. Whether it's getting stabbed in the neck by some zombie that your taxpayer dollars are paying to keep drugged up, a foreign truck driver with a fake CDL issued by your state DMV plowing through your family minivan, or the esophageal cancer you get from the carcinogenic lining in the save-the-turtles paper straws they make you use...despite all the Chicken Littles throughout time, warning of some cinematic end to your life, it's communists in government who've quietly racked up 100M+ body count in the last century. Stop indulging these cartoonish macabre fantasies and get real.

Alexander PCs avatar
Alexander PCs@AlexanderPCS·Nov 13

Is Rick Cerrone on X? Because this is not a defensible MVP ballot. Good lord. Someone take this guy’s voting privileges away.

I would like for you all to look at that 2nd Pittsburgh voter’s ballot because oh my lord

Post attachment
will depue avatar
⟳ Alexander PCs retweeted
will depue@willdepue·Nov 12

I have a major result to share. I’ve found conclusive forensic evidence that the top 4 Chinese AI labs—DeepSeek, Moonshot, MiniMax and Ant Group—have built their newest models on NVIDIA’s most advanced Blackwell chips, which are barred from export to China. This sounds insane, but I discovered that almost every AI model carries a hidden signature of the GPU that trained it. When a model is first set up, the starting values of its weights are laid out in a pattern that depends on the exact GPU. Each combination of GPU and seed number produces a unique pattern—a sort of fingerprint of the hardware used to create the model. Traces of that pattern survive training into the published model. Again, this evidence isn’t based on any insider information or leaking. It’s based on a forensic analysis of the model weights released by the companies themselves. I searched more than six trillion combinations of chip layout and seed number and matched them to open-weights models. In doing so, I found conclusive evidence that six Chinese AI models—including DeepSeek-V4-Pro, Kimi K3, and MiniMax-M3—were initialized on NVIDIA Blackwell systems. Four of the models match Grace Blackwell, the GB200 and GB300 parts in NVIDIA’s flagship liquid-cooled rack-scale systems. I then verified these results on real hardware. Running unmodified PyTorch with the seeds I’d recovered, a Blackwell B300 reproduced DeepSeek-V4-Flash’s untouched starting weights bit for bit, and a rented GB200 reproduced Kimi K3’s to a correlation of 0.999995. A Hopper H100 reproduced none of DeepSeek-V4-Flash’s, and changing the seed by one broke the match. (DeepSeek’s seed is 953049583 by the way.) The method also was verified to match 24 known open-weights models to the hardware that developers said they used. It is exceedingly unlikely that these results are a coincidence. Across six trillion combinations, the best accidental match should reach only about 8 sigma (i.e., eight standard deviations above what random chance would produce). The real matches range from 49 to 1,385 sigma, and even the weakest single test hits 20. These were data-center-scale training runs. With some analysis of release schedules and model details, I estimate that the largest, Kimi K3 and DeepSeek-V4-Pro, used 2,000 to 3,000 Blackwell GPUs each. Clusters of this size cost on the order of $150 to $200 million each to buy, and across these labs alone would add up to more than half a billion dollars of export-controlled GPUs, not counting experimental compute or otherwise. It’s quite likely that we see multi-billion dollar expenditure on export-controlled GPUs. I am highly confident in this result, and it has been independently verified by senior researchers at OpenAI and Anthropic. You can read the full report at unemploymentlabs.com/checkpo…. There is much more I’ve identified about these models through other forms of checkpoint forensics like this, and will have more to share soon.

Post attachment
joshpuckett avatar
⟳ Alexander PCs retweeted
joshpuckett@joshpuckett·Nov 12

I've been surprised to learn that so many of you believe actions should not have consequences, and leaders should not be held to a higher standard.