Christian Antkow (@CAntkow) header
Christian Antkow (@CAntkow) profile picture

Retired: Former id Software, Oculus/Facebook Hermit Rancher & Farm Machinery Operator DGX Spark Enjoyer and Open Weights Connoisseur. LLM Promptineer.

Citadel 21 Ranch, Texas
Joined July 2017
Photo by Christian Antkow (@CAntkow)Photo by Christian Antkow (@CAntkow)Photo by Christian Antkow (@CAntkow)Photo by Christian Antkow (@CAntkow)Photo by Christian Antkow (@CAntkow)Photo by Christian Antkow (@CAntkow)Photo by Christian Antkow (@CAntkow)Photo by Christian Antkow (@CAntkow)Photo by Christian Antkow (@CAntkow)Photo by Christian Antkow (@CAntkow)
So, I saw that dude used AI to clean-room reproduce all the Adobe products for Mac, and they kinda don't entirely suck Was really bummed to see that nobody did an Adobe Audition clean-room version. Kinda took it personally With AI, I can just do things github.com/BobClawblaw/spect…
1
2
101
Christian Antkow retweeted
Working at Oculus would have been an utter waste of my time except for one thing. While I was there, I realized what's wrong with leftists. Not that they have been programmed, that's obvious. Not why they have been programmed. For power, that's obvious, too. But how they have been programmed. The exact method that their moral compass has been destroyed works like this: Target a bunch of people with very few deeds and accomplishments, such as young people or incompetent people, or incompetent young people. Create for them an environment where they are rewarded with moral approval, not for what they do, but for believing certain ideas. The ideas don't have to be outrageous, or absurd, or have anything to do with what you actually need them to believe. In fact, it's better to start with something innocuous. Why? Because the goal isn't to install those beliefs at all. The goal is to repeat this process until you create a person who measures his moral worth by what he believes, not what he does. Keep at this until you have created a mass of people who not only reward themselves with a feeling of moral superiority for believing the right things, but who will reward each other for believing orthodox things, and who will compete with each other, not to do the most good, but to have the most orthodox beliefs. Now, change one of the propositions you have programmed them with to a slightly more radical version of itself, and hold this up as a more correct version of the old belief, which has been newly discovered by scholarly investigations into culture and moral philosophy. Sit back, watch, and give them a few weeks to compete with each other to adopt the new belief, and to condemn the "evil" of anyone who believes the old belief, the very thing they believed last month. Repeat this process until they are used to rapidly changing their entire belief system in response to an external cue. Congratulations, you have a population of entirely programmable people whose moral compass is not just broken, but entirely non-existent. They don't care what any person in front of them actually does. Only that he believe the orthodox thing, which they just learned about in last week's memo. They are willing to do anything, so long as it is done in support of their current belief system, which can be replaced whenever it's convenient to do so. They are willing to burn down an entire city with no qualms whatsoever if they are doing it in support of the idea that black lives matter. They are willing to murder a man in front of his children, and it doesn't matter if his actions were entirely peaceful and good, because all that matters to them is whether he believes the correct things or not. And a man who builds cheap spaceflight for the human race, develops viable electric cars to save the environment, restores voices to the voiceless, and plans, for his next trick, to heal the blind... well, all those are deeds, so none of them matter. The only thing that matters is that he told people to vote republican once, which is wrongthink. So what you have a horde of communist zombies, who can be steered to shuffle in any direction and attack the living, eat their brains, and turn them into more zombies. Once someone becomes one of the zombies, you can't cure him, because a moral compass can only be repaired by understanding the consequences of actions in a new way... and they don't care about actions or consequences. They only care about beliefs, and not about consequences at all. Their moral compasses aren't just broken. They are absent. There's nothing to repair. You have to understand that the zombies are zombies now.
The amount of pro-Elon bots on this site are ridiculous. Ain’t no way this many people support this literal comic book supervillain.
445
2,335
313
10,605
207,228
Christian Antkow retweeted
Jim Cramer is a legend
593
3,862
757
26,491
1,981,306
Christian Antkow retweeted
House of cards now falling apart *OPENAI SLASHES LATEST FUNDING TARGET BY $10B: FT *OPENAI NOW SEEKING ABOUT $20B AT A $1.25T VALUATION: FT Translation: Abu Dhabi passed on $30BN.
199
557
92
7,239
760,008
Christian Antkow retweeted
I am hesitant to weigh in on discussions about 30+ year old details. I understand enough about memory in general, and mine in particular, to not have great confidence about it. I generally trust Romero’s memory more than mine — I was always a bit jealous of that capability. In broad terms, multiplayer gaming was right up there with 3D gaming in my grand visions of what I wanted to accomplish. The early FPS games were very much “first person Gauntlet”, and Gauntlet had multiplayer. I do clearly remember the first time I got two computers running Doom taking to each other. Moving around on one computer and seeing the character move on the other screen was sublime. And then, of course, I shot the other character. Every shared virtual environment inhabited by young men inevitably hosts deathmatch. I suspect the confusion here may be around the fraction of people that originally bought Doom because of multiplayer. It was almost certainly well under 10%. Maybe under 1%. It wrecked some college networks, but most people didn’t have Ethernet cards or null modem cables. We probably did have some conversations correctly predicting that. Deathmatch was awesome, but most of our players never got to experience it. Even four games later for Quake 3 Arena, being “deathmatch focused” was a huge internal battle, and not at all obviously viable.
Replying to @romero
John. First, 8 months before release is before I joined the company. So I have no knowledge of what you did then. second, AFTER I joined we had a discussion in which I talked about my experience with Command HQ and YOU YOURSELF agreed that Deathmatch wouldn’t be that big a deal. Do you not remember? Third yes we all wanted Deathmatch anyway. And I’m glad my predictions were wrong. Perhaps check your own memory before gaslighting mine. Hope you’re well.
219
241
46
5,019
485,600
It's a tough life, when the hardest computing decision I need to make locally, is whether I use Deepseek v4.1 Flash 1017 or GLM 5.3 Flash for my local workload. What an amazing time to be alive.
1
1
61
Christian Antkow retweeted
Replying to @SandyofCthulhu
Hi Sandy. Are you well? This is wrong. On January 1, 1993, we sent a press release saying that multiplayer would be in DOOM, a full 8 months before you joined the company. We knew deathmatch would be huge. In that press release, in January 1993, we said multiplayer would be "the number one cause of decreased productivity in businesses around the world." We were sure of it, and we needed no convincing. I understand you are proud of your work on DOOM, and you should be, but please also respect the accuracy of your statements, particularly around game history. My concern is that these false statements are designed to trigger replies (like this one), reposts, and outrage discussion, all with the view of increasing your followers and subscriptions, particularly since this false deathmatch story (among others) is so easily disproved. I’ve even welcomed you to reach out to me ahead of time. You seem to appreciate game history as much as I do. Could you not also make yours accurate?
223
1,119
251
20,824
1,144,426
The Circle of Life Caught this little dude chomping away on a grasshopper as I was leaving the ranch yesterday.
1
2
61
Christian Antkow retweeted
I don’t always drink beer, but when I do, it’s to Romero pwning Sandy on Twitter.
Replying to @SandyofCthulhu
Hi Sandy. Are you well? This is wrong. On January 1, 1993, we sent a press release saying that multiplayer would be in DOOM, a full 8 months before you joined the company. We knew deathmatch would be huge. In that press release, in January 1993, we said multiplayer would be "the number one cause of decreased productivity in businesses around the world." We were sure of it, and we needed no convincing. I understand you are proud of your work on DOOM, and you should be, but please also respect the accuracy of your statements, particularly around game history. My concern is that these false statements are designed to trigger replies (like this one), reposts, and outrage discussion, all with the view of increasing your followers and subscriptions, particularly since this false deathmatch story (among others) is so easily disproved. I’ve even welcomed you to reach out to me ahead of time. You seem to appreciate game history as much as I do. Could you not also make yours accurate?
17
19
448
16,187
Well, it's technically end of October. Wonder if Apple is going to offer their 512gb Studio M5 Ultra's today. Want to buy two. Going really deep down this whole local AI rabbit-hole and having so much fun!!! What a time to be alive.
1
1
2
69
This weekend, I'm having a local abliterated Deepseek 4.1 Flash 1017 model running across two DGX Sparks beating on my IOT devices, with a goal to root our Samsung television. What could possibly go wrong? What a crazy time to be alive. You can just do things.
1
3
111
Christian Antkow retweeted
Finally full DOOM on a Vape😅 Source: github.com/atc1441/Vape_DOOM… Since the previous custom firmware was only Screen Sharing it this was finally needed! Only possible by Arturs impressive work on making DOOM go bzzz on 32kb RAM
101
1,105
153
9,715
551,185
Abliterated local LLMs and agent harnesses are one of the coolest developments I've yet experienced in this lifetime. What an amazing time to be alive.
1
29
Uhh, uhhm... How does one get an invite for the Tesla Roadster Rocket pre-order ?
4
178
Christian Antkow retweeted
Whenever I see a decorative mirror ball, I think about the old-school OpenGL GL_SPHERE_MAP texgen option, and light probes for realistic lighting of synthetic renderings for merging with photos. Long before cube maps were an option, sphere maps allowed a form of shiny surface rendering, but only from a limited viewpoint — you couldn’t walk all the way around a shiny object. For light probes, you could take multiple photos with different exposure times to capture a true many-orders-of-magnitude high dynamic range composite image for realistic image based lighting of everything from direct sunlight to subtle shadowed region indirect lighting.
77
77
2
1,378
90,860
It's only a Tuesday, and I've already blown through half my Claude Fable 5.1 allocation for the week. OTOH, I've really been learning to love GLM 5.3 Flash locally when I need to do serious work.
1
40
Christian Antkow retweeted
Very few will care about the actual reinforcement learning details here, but watching me grope around for understanding in the chat log might be of broader interest: chatgpt.com/share/6aad4252-f… I have often noticed that our estimated Q values are higher than the observed returns, sometimes substantially so, and that can’t be good for performance. For policy decisions, only the relative values matter, but an offset negatively impacts bootstrapping, so I was excited to see this paper on Relative Value Learning: arxiv.org/pdf/1901.09732 After struggling a bit with “the Banach space of bounded antisymmetric pairwise functions“, I realized all that is essentially going on is subtracting two value functions. Given this framing, Chat was able to simplify the machinery into a particularly elegant form: Subtracting the mean TD error from the individual sample TD errors gives all the benefits of relative value learning. Summarized as "relative Bellman regression is TD-error centering". Unfortunately, while you can create examples where this should be very valuable (someone should write a proper paper on it!), it didn’t actually improve performance on my tasks. However, I think this has usefully narrowed down what is actually happening. Bellman iteration naturally corrects towards a correct absolute value, but only when the bootstrap values are also training targets. With Q-learning, you are often / mostly bootstrapping from a max-action that was not actually taken, so it never sees any downward Bellman pressure. This is consistent with another result I have noted: learning state value functions offline from a frozen policy without actions doesn’t seem to suffer from any value overestimation.
61
36
10
553
125,877
Christian Antkow retweeted
It is crazy that the People's Republic of China has sanctioned me as a seperatist terrorist, but even crazier that so many American executives and influencers still shill for them. "How could anyone have predicted this outcome?", they will cry as sanctions inevitably hit them.
Replying to @PalmerLuckey
中國制裁我的原因:Altius-600M 在台灣海上目標射擊百發百中. teapaw.org/anduriltech/status/209…
416
1,226
69
15,880
704,419
... and @EmbarkStudios @ARCRaidersGame still won't add PvE mode. I've never witnessed a company repeatedly shoot itself in the foot as much as Embark have. They just can't learn.
42