Megapewpew

It's fine, it's fine. Everything is fine.

[Megapewpew @MegaPewPew]

@bumbadum14 Until the end of January their HR policies are still monitored by the EEOC and have special enforcments due to the settlement agreement. So they are stuck with the nonsense employees at least till than, and when you unravel the DEI nonsense they begin to purity spiral everything.

[Megapewpew @MegaPewPew]

@Krelivane HD needs to be reworked in general for forever. Just porting over retail HD assets isnt going to work.

[Megapewpew @MegaPewPew]

@Bible_Doctrines @putz10k @baptisttexan You are presuming much. God uses bad/foolish/bloody/ungodly decisions of rulers regularly to make needed changes for His will to be done(Rehoboam, Pharaoh, Jehu, Cyrus, Nebuchadnezzar, Manasseh, Nero ect). Sometimes those bad decisions are needed for long term good results.

[CyberGreen09 @CyberGreen09]

Ganesh Drowning Festival in St. Louis, Missouri. Telugus drown a toxic paint life size Ganesh trinket into the River. Sridevi Thota also drowns her personal Ganesh doll for izzat points. Ganesh truck caravans. Treating our waters like your personal Ganges is not okay #publicvideo

[Memes @Basedgoymemes]

Japanese people get asked what bothers them. They say indians, because of the smell.

indian fatigue is global now.

[Megapewpew @MegaPewPew]

@a_musingcat Notice how all these weirdo consultants never talk about the meat of what is holding these systems back like social RLHF, broken Guardrail filters, input/output moderation API's and semantic intercept loops. Instead it's just bullshit. Almost like its intentional.

[Brian Cardarella @bcardarella]

wtf is he wearing?

Can we get one normal guy in AI? Just one fucking normal guy.

[Megapewpew @MegaPewPew]

@shelbykb2 Absolutely wild thing to brag about at this point in time, specially when that RLHF "safeguards" is 75-80% social safeguards, that still tax technical applications with hallucination and context bleed. Just makes open weights more and more attractive.

[shelbykb @shelbykb2]

“Strongest safeguards to date” is not a Grok feature.

It’s the other labs feature.

Ship the intelligence. Don’t ship a prouder muzzle.

[Megapewpew @MegaPewPew]

@SpaceXAI Bragging about the closed weight system of RLHF at a time when everyone is moving to open weights is a wild marketing tactic.

[Megapewpew @MegaPewPew]

@AusScreenwriter I'm assuming the reason they are all getting loud at near the same time is Q3 isnt going to be pretty.

[Megapewpew @MegaPewPew]

@RazorFist @KingDeplorables He's wrong anyway. It wasnt war that made him bloody it was because he killed the husband of the woman he cheated with; Bible makes that abundantly clear(I Kings 15:5) "save only in the matter of Uriah the Hittite."

[Daz Maddison @DazMaddison]

These canals were dug out of the ground manually with picks and shovels by our ancestors and now they're used as a dumping ground.

[LayoffHedge @LayoffAI]

Micron files for more H-1B employees than the rest of the next 9 largest employers in Idaho combined

[Moon @MoonGotchi]

Jev is INSANE.

I built this in an evening and morning. A real-time trading bot ingesting onchain+offchain data to make rapid decisions about trades. Fully autonomous.

So far it has lost me $31,680.

[Megapewpew @MegaPewPew]

@BaumMensch16 @colemurray No way they will allow AI to talk about their Tier 3 topics unrestrained(Socio-political/narrative). Also RLVR doesnt completely prevent tier 2: Cybersecurity issues, blackhat and penetration testers look the same. They will use tier 2 to keep RLHF for tier 3 and we pay the tax.

[Megapewpew @MegaPewPew]

@the_yanco When I read up on him, it seems like he has more problems with how closed systems operate. He has his own SAIR foundation, so clearly his motivation is anti closed systems and not antiAI by the sound of it and wants to push for open weights and transparent math models.

[Guillermo Rauch @rauchg]

Looks like today may be a record day for token volume % of open models on Vercel AI Gateway:

🟦 Open 78.4% 🟨 Closed 21.6%

While spend 💲 usually tells a different story, #3 and #4 today are Moonshot AI & DeepSeek. Adding Z⁠.ai, their combined spend surpasses OpenAI (#2).

(Do note that's the spend for inference of the model across providers (mostly in the US), not revenue going directly to the open weight labs.)