Arman Tarkhanian

Professional LLM engineer. My opinion doesn't represent that of any organization.

[IGN @IGN]

ConcernedApe's Haunted Chocolatier is "more of an action RPG" and "it's for the people in Stardew Valley that go into the mines." https://t.co/q1BQ1gEIPA

[morgan ♡ @picmitsuba]

what confuses me about discussions of this dress or anything adjacent is that its so blatantly in a light environment. theres no way it could be in the shade for it to be white rather than blue 😭 you can see an actual white and black dress right behind it 😭😭😭😭

[Arman Tarkhanian @armantark]

@JianOuyang5 I want unlimited chatting inference tho 😩

[Arman Tarkhanian @armantark]

These 28 days of improvements from OpenAI have been lackluster so far. As expected. The best thing is Intelligent UI but that's still too minor.
1. GPT-6 "Sol" (which is actually just 6 Terra) hallucinates way too much and is extremely lazy even on Extra High but you can't use Intelligent UI on Pro for some reason?? Just give us Astra or 6.1 Sol bro. And there's still the weekly limit for Pro. Not that I've gotten to it but still. 6 "Sol" is probably worse than 5.6 Sol.
2. We really need skills for ChatGPT. The enterprise/business/education accounts get them but not consumers for some reason. I'd be happy if they just threw it in one of these 28 days.
3. Dot, like everyone else says, is flatout underwhelming, and I'm only using it because it's free and a better model than Grok or Muse. I know Grokbot is starting to use Opus 5.5 but it's still quota based and I'm on the meager $20 plan.

[lauren @poteto]

i feel like swe interviews could probably collapse into just 2 technical rounds:

• system design (how well can you articulate your ideas, do you actually know how to engineer something)
• onsite project to build a real thing with agents (can you translate your intent into high quality outcomes)

everything else we used to ask is no longer necessary

[Arman Tarkhanian @armantark]

Who tf actually likes this type of feature? It's just annoying and always gets it wrong

[OpenAI Developers @OpenAIDevs]

Now in beta: composer predictions in Codex for Pro users.

Codex can now suggest your next message based on your conversation and how you talk to it.

One of the most loved new features we've ever tested internally.

[Arman Tarkhanian @armantark]

@Rasalas8910 @theo Sometimes I want to add a feature to an archived project which is in a random folder, but that does make sense...

[Arman Tarkhanian @armantark]

I don't even have autism and I've heard this before post-interview. I just speak monotonously, jeez

[Arman Tarkhanian @armantark]

Waiting for this to mature a bit, but this sounds really promising! Excited.

[Taelin @VictorTaelin]

My harness: https://t.co/AsN1MY2e1g

This is the closest to infinite context I've experienced.

I've been using it for a week to organize my life and build things (including itself!), and it just keeps going. I only use it now.

OSS next week, prompt to build your own below ↓

[Micheál Ganley @GanleyMicheal]

Wrong. ADHD is a superpower. You just must embrace being a generalist.

E.g., Charles M. Schwab, CEO of Bethlehem Steel (and instrumental in building the arsenal that won two world wars), clearly maximized his ADHD:

[Arman Tarkhanian @armantark]

About frickin time bro. Unless it's Luna 😩

[Tibo @thsottiaux]

Day 3/

The big one is GPT-6 in Chat, but today is also a little celebration day with a new high of 40M active users across Codex and ChatGPT Work.

Loading a banked reset in everyone's paid accounts. See you again tomorrow!

[Arman Tarkhanian @armantark]

As an ADHDer myself, uhhh, yeah, infantilizing yourself is not great.

[MJ @morganisawizard]

the tweet, btw

[Arman Tarkhanian @armantark]

How does Theo have this much pull

[Elon Musk @elonmusk]

Important note regarding Grok @Bot:

Going forward, @SpaceX will use the best back end model for any given task, including Claude Opus 5.5, MidJourney, Suno and other leading APIs.

Whatever is most likely to give you the best outcome.

[Proctor Zakharov @ProctorZ]

Oh boy, it's time to get out the baseball bat again and really pound the balls of someone with a worthless opinion in the modding scene.

A.I decompilation is the best thing that ever happened to gaming. It has, overnight, eliminated the predatory stranglehold that game developers have tried to inflict on the marketplace for decades. It has revitalized dead modding scenes and overcome issues that had stumped private projects for years. This is unambiguously a good thing.

Yet modders like this man are mad and raging incoherently about it. Why is that? Simple answer: it takes away their power.

I'm in a much better position to say this than most, because before I was a vtuber fan and got a real life job, I was a modder. I worked on various simple homebrew projects and contributed to a number of mods that I was interested in. My only coding experience comes from banging my head against NWscript long enough to make sense of it purely via observational learning.

During my time in the modding scene, I was able to observe first hand just what paranoid, tinpot dictators most modders actually were. They were selfish, vain, miserable, mercurial, prone to spontaneous bouts of megalomania, and often abandoned their projects at the first sign of critique or negative feedback.

To this day I quietly mourn the many projects I watched be sabotaged by entitled modders who would rather destroy everything they had worked on than pass it to another person who might change their precious, sacred vision of it. There are wonderful stories and creative writing projects I sunk days of my life into that are dead because one entitled little code-gremlin threw a hissy fit and went on a rampage.

That's why I say you should hate modders as much as you should hate journalists. There are some that are good people, but most are tyrants and egotistical psychopaths who only care about modding because it is a way for them to exert power over a community and get attention when everything else in their life is worthless.

The democratization of modding tools by A.I is unambiguously a good thing. I hope it continues and gets bigger. Every time I see a post like this, a little part of me is happy, because it means one of the vile rats who ruined my enjoyment of the modding scene for the games I loved back in the late 00's and early 10's is suffering the completely deserved fate of utter irrelevancy.

[Arman Tarkhanian @armantark]

Still, I do miss those days... It felt like an esoteric art to even know what "chain of thought prompting" was to make an LLM anything more than a simple toy. Now it's just normal.

[Theo Jaffee @theojaffee]

Yeah I think it's safe to say at this point that "prompt engineers" of the 2022-24 variety were wrong. Prompting is fundamentally just giving the model the right context to do what it needs to do, which is a basic communication skill. No gimmicks needed

[Tenobrus @tenobrus]

astra ultrafast can decompile and natively port a windows steam game to mac in about 2 hours

[Arman Tarkhanian @armantark]

And I say this as someone who does in fact value the "soul" of art, despite my professional job title. There is something inherently unique to human consciousness that we can never replicate.

[Arman Tarkhanian @armantark]

Genuinely do not understand this rhetoric. Isn't the whole point of AI to reduce menial labor? Decompiling a game is the definition of stuff that should be outsourced to a computer. It requires no human creativity or "soul" required. It's just pattern matching. Which, sure, I guess, if you like the puzzle of it.

[JD | RoyalCities @RoyalCities]

New AI hate meta just dropped.

Gaming recomps have a soul now and you should feel bad for playing anything that didn’t take 5 years to make with handmade, free-range, artisanal code.