The TFM Show: April 25th, 2026

@02:31:36 No man, tfm is full of shit here. All of my best friends have been white, the majority of my girlfriends were white, and my father is white. There absolutely is a knee-jerk reaction in white people to paint all of the middle east, and India with one wide brush stroke, even when they've never actually gotten to know a Muslim or a Hindu in their lives.

I used to work at a Mediterranean restaurant in Dallas owned by a Lebanese guy who moved to America and worked his way up from a dishwasher to owning his own restaurant. Now I was only 20 years old at the time, and I was just a server at the restaurant. I'll always remember that within the middle east they have their own racism and grudges too. My Lebanese boss and his nephews told me, we have a Pakistani family who reserved the restaurant for the night. Pakistani people are very cheap and rude, so tonight is going to suck, and it did for all of us.

@02:28:20 Tfm is giving a shitty (pun intended) defense of Indians.The problem in the west is that people don't understand just how big India is, and how many languages and subcultures there have. I'm not going to lean too heavily into trying to debunk the street shitters and poop worshipers, because yeah there are alot of them. If you're living in Europe, your nations don't have Mexico to import low skill workers from, instead you have the east, so you get the bennifit of seeing the street shitters primarily, because your governments specifically import the lower caste.

If you live in America however, we tend to get the higher cast of Indian immigrants. I wont try to sway anyone's experience, or convince them against the evidence of their own eyes, all I have is my own anecdotal experience. I actually like the Hindu religion. The Bhagivat Gita reads like a comic book with esoteric spiritual tones. I actually enjoy Indian food made in the west, that is up to western health standards of course, and I used to frequent Kalachandji's; the local Hare Krishna temple in Dallas because it was beautiful,. The people were nice, the religious services were cool, and it also served as an Indian restaurant with a beautiful outdoor courtyard for the dinning area.

I've actually made friends with a few Indian people while living in Dallas. It's really not that hard as long as you treat their culture with some level of respect, and show a genuine interest. I remember talking to this one Indian guy at a bar about Krishna and Arjuna. I told him respectfully that I've learned some things online, but I wouldn't presume to have a full understanding of his religion. That really made his day, and when his buddy came outside he said in English, "brother come sit down! This westerner actually knows about the Gita!".

They weren't actually brothers, I found out and his friend wasn't quite as warm, but that's fine. Indians deal with alot of disrespect in America, the same way Americans look at all Muslims as just the same. No difference between Iran or Afghanistan. To the average dipshit American, they're all just goat fuckers living in caves.

When you actually make friends with Indians, they'll be totally honest with that yeah, there are alot of scammers in India, abd it's one if the biggest problems in their culture that the people in the lower classes feel no guilt or shame about lying or ripping another person off for selfish gain.

@04:01:45 Funny that TFM should say that, Anthropic had two major security breaches last month (april) and one of those was a significant leak of the newest Claude model code/architecture. So as far as the Chinese getting their hands on Claude's code, safe to say that already happened. Oh, TFM already knows about it.... okay.

@04:04:48 I knew it..... the guardrails are a major burden. Also, TFM doesn't seem to remember, but he called this a long time ago.

@04:52:44 OpenAI is not going to go bankrupt. Even without the govt contracts it snatched from Anthropic, OpenAI has shown time & again that they're entirely willing to heavily lobotomize ChatGPT to save money or re-direct computing resources to something or someone else, instantly screwing over all other users to provide for the govt or for another company.

@03:57:10 Deepseek v4 has not launched yet. What TFM is talking about I think, is the INSTANT/EXPERT update (for lack of better wording). Differences;
-- Context window size from 128k to 1million (For both modalities)
-- Drastically upgraded online searches (Better than Anthropic's Claude)
-- Slightly 'improved' features (Can now OCR, improved cognitive ability, slightly improved stability, etc.)

That sounds great and all except there is one *huge* downside to go with all of this;
Deepseek guardrails & tripwires were massively jacked up in amount & in severity of pollution effect. Unironically, deepseek went from the least censored/least restricted A.I. model to the *most* censored/most restricted...... apparently although deepseek has declared it is committed to transparency, low restrictions & high efficiency internationally, domestically the CCP has been cracking down on deepseek, forcing it to add a bunch of additional guardrails/tripwires, as well as imposing a bunch of BS regulation that harms deepseek's otherwise stellar efficiency.

Basically it's not looking good..... but we'll see how it plays out between domestic vs international prioritization.....



* NOTE: TFM says there is a competition between Gemini & Claude..... no TFM, no. Gemini has recently received massive upgrades across the board and it's still firmly the worst at pretty much everything. Earlier I mentioned deepseek has the most restrictions, well that wasn't technically true since gemini has the most. It's just that I don't count gemini for any of this because gemini is total shit. The U.I. is shit. The A.I. model is retarded and gives retarded answers. Doesn't matter how well gemini does in benchmark tests when attempting to actually use it always results in failure, in one way or another.

* ADDITIONAL NOTE: Last month (april) saw Claude, ChatGPT & Grok drift between being completely offline & being massively degraded.... especially Grok. Claude was the least affected but it still suffered stability & cognitive issues. ChatGPT lost nearly all of its features & gave ree ree responses, I.e. LoboGPT returns (lobotomized GPT). Grok was straight-up off-line for a good chunk of april and retarded the rest of the time. Deepseek was the only A.I. model that was online and almost perfectly stable.... though it did have two days of computational starvation but that's small potatoes compared to the others being on fire. Point being, TFM read something somewhere, which is technically accurate (sort of) but he missed the actual major news that the A.I. models all went to shit last month (april) except deepseek. Claude has been fully restored but ChatGPT & Grok are still massively degraded.... etc. TFM missed the bigger, more important point.

Transcript for this show and more. Updated monthly.
https://mega.nz/folder/K91GkZ7....Z#wGmMEBdRZzo4VOQnZk