Elon Musk pushes out more xAI founders as AI coding effort falters

All: please stick to thoughtful, substantive discussion. You may not owe you-know-whom better, but you owe this community better if you're participating in it.

If you don't have a thoughtful, substantive comment to add, not commenting is also a good option. There are quite a few interesting submissions to talk about.

https://news.ycombinator.com/newsguidelines.html

I think the problem for xAI is that it can really only hire two types of researchers - people who are philosophically aligned with Elon, and people who are solely money-motivated (not a judgment). But frontier AI research is a field with a lot of top talent who have strong philosophical motivation for their work, and those philosophies are often completely at odds with Elon. OpenAI and Anthropic have philosophical niches that are much better at attracting the current cream of the crop, and I don't really see how xAI can compete with that.

Feel like the canary was when Grokpedia became a project.

Giant waste of time while Anthropic/OAI keep surging forward.

I also keep hearing this narrative that Twitter is a good data source, but I cannot imagine it's a valuable dataset. Sure keeping up with realtime topics can be useful, but I am not sure how much of a product that is.

I feel xAI is just a very big version of the Boring Co. "flamethrower": an unserious endeavor which is just a reskinned existing tool (it was a reskinned weed burner), but people were wowed by it anyway, since Musk was behind it, and they all pretended it was something new and notable.

The burning (heh) question is which SpaceX subsidiary will fail first, xAI or Tesla (not yet a subsidiary, but it's written in the stars (heh))?

Then again SpaceX is also jumping the shark what with their orbital data centers (remember those?).

Might be time to start a new Musk company soon.

I don't use it myself, but I feel like the way Grok is integrated into Twitter is a pretty good thing for discussions, as it is certainly a more objective and rational voice than most human participants. I think it's good that people tag @grok if they don't understand something or want an opinion, even if it looks pretty silly to see "@grok is this true" repeated multiple times in replies.

That said, Musk's attempts at misaligning the thing and make it prefer his opinions of course destroy any trust. It's surprising that it's seemingly as good and helpful as it is despite the corruption attempts.

I also don't quite get how the business model is supposed to work out if its main usecase is to serve Twitter. I know they provide API access as all other models, but with how distrusted Musk is and how sensitive of a topic reliable model behavior is, they seem to sabotage themselves. Which company wants it to go mechahitler on them?

Used Grok for the first time, in a Tesla, and for that purpose it actually made a lot of sense. It’s very well-integrated into the car’s systems and communication style while driving tends to be very tweet-esque. I think this is the niche they should lean into more (live assistant, e.g. Jarvis type stuff) and leave the more agentic niche to folks like Anthropic. Maybe even delegate more difficult or background tasks to those sorts of models. As a verbal interface I found it pretty pleasant.

While I believe Grok was a decent model (in some of our internal use cases it performed the best until Gemini 2.5-pro came out), I can't help lament how the team chose to run.

xAI (and Twitter) was the loudest about six-hour workdays, sleeping in the office, and always shipping. ~2 years later it feels like they have nothing to show for it. I'm sure the engineers at Google worked 4 days a week, 2 hours a day, with half of that being spent at the Google cafeteria and they dusted xAI years ago.

The irony is that while Wikipedia faces criticism for bias, it remains one of the few massive-scale sites with a clean internal link structure that doesn't feel manipulated by modern SEO 'clustering' tactics. For developers, their API is still a masterclass in how to serve structured data to the public.

Recent, related, and apparently ahead of the curve:

Ask HN: What Happened to xAI? - https://news.ycombinator.com/item?id=47323236 - March 2026 (6 comments)

“Orbital space centres and mass drivers on the Moon will be incredible.” - Musk

Right.

The product is the stock. TSLA: [1] Up by 3x in the last two years, despite no new models, the Cybertruck failure, the Robotaxi failure, the large truck failure, and an overall decline in sales. How does he do it?

It's a concern seeing Space-X, which builds good rockets, drawn into the X and AI money drains. Space-X is needed. If X and X/AI tanked, nobody would care.

[1] https://www.cnbc.com/quotes/TSLA

xAI's biggest contribution to the space seems to have been their x-rated image/video model. Hard to see what xAI has to offer against Gemini, Claud, ChatGPT.

Nobody wants a biased AI like xAI. If I want a biased opinion, I can just ask my neighbor. A good AI feels like the collective knowledge of humanity at my fingertips, not the random ramblings of a lonely old man.

> "AI was not built right first time around, so is being rebuilt from the foundations up"

So Tesla's recent $2 billion investment in xAI was a bad deal?

It looks a lot like a public company is being used to bail out a private one.

This is veiled speak for "No one wants to work for us, so we need to contact rejected applicants to fill positions".

I use AI for work, but not agentic, at most per method/function using GitHub CoPilot (which has Grok on it).

Grok is at best useful for commenting code.

> Toby Pohlen, a former DeepMind researcher, was put in charge of the “Macrohard” project to build digital agents that Musk said could replicate entire software companies. Musk said it was the “most important” drive at the company. The name is a “funny” reference to Microsoft, the billionaire added. Pohlen left 16 days later.

When I was 9 years old, my uncle asked me what I was going to do for work when I got older. I told him I was going to start a company called "MacroHard", and become the richest man alive. He told me that's not how the world works. Turns out it is.

> Recruiters have been contacting unsuccessful candidates from previous interviews and assessments to offer them jobs, often on better financial terms, the people said.

I'm not sure those candidates would want to work for xAI after seeing the news and everything unless they desperately need a job right now.

It's not hard to imagine getting laid off or fired weeks if not days after joining the company.

Grok's UVP is still nonconsensual porn, right?

Maybe they shouldn't have spent so much time trying to make their model have an edgy cringe attitude, Idk.

Obviously catching up to others in agent assisted coding is the motivation for this. But it is also an odd decision in the same way that Meta hiring an AI leader from a data labeling company is odd.

Grok 4.20-beta1 scores above GPT-5.4-high and just behind Opus 4.6 on LMArena for Text https://arena.ai/leaderboard

I guess for coding if you’re not first you’re last, but this is damn impressive considering. It looked like they pulled the coding model from the benchmarks, but it was similar.

It feels like xAI is perpetually playing catch-up.

They haven't quite committed enough to a novel direction relative to anthropic or OAI, what's described in the OP seems symptomatic of a lack of differentiation.

If you spend all your time judging yourself relative to the incumbents, there will be no time left over to innovate.

The leash is too tight!

I've been saying this for a while, but if I had to use Grok for anything programming-related I'd feel very sad and unproductive. I was playing around with a local TTS model codebase but having some issues getting it to work, so I tried explaining the problem to all the major models to see how they performed. Grok performed the worst by a significant margin, and the worst part was that it easily became stuck trying minor changes that didn't solve the key problem.

If we are to take any claims of Recursive Self Improvement seriously at all, then having a competent coding model seems like a key asset where you need to guarantee that you're remaining competitive. Why wouldn't you make coding models a top priority if you expect it to ultimately help your internal teams become more productive and effective?

There's also not an unlimited supply of researchers and engineers for them to keep burning through people at the rate at which they've been working. Although I guess for people with short timelines it makes sense to sprint hard, while people with longer timelines are more likely to treat this as a marathon. Maybe the years of burning bridges and developing such a toxic reputation are finally catching up to Elon. I think part of the harm that Elon has done is framing all the work in xAI as engineering while being highly dismissive of research, but a lot of research requires running experiments or thinking about problems and exploring them for long periods of time. If you're just grinding out work nonstop you don't really have time to let your mind wander and explore new ideas.

Honestly, I'm surprised they've done such a terrible job with programming. I remember around summer last year it was quite apparent how far behind they were with coding tools, but Elon was posting about taking that domain a bit more seriously. Why didn't any of those efforts materialize into real outputs? Something must be truly dysfunctional inside of xAI for them not to be shipping anything at all, especially considering Elon's propensity to ship undercooked products while continuing to iterate on them, as he has done in many previous cases.

I've noticed that Elon has also gone very hard on social media posting a ton of criticisms against the other big AI company CEOs like Daario Amodei. This suggests to me that he must feel very threatened, otherwise he wouldn't be resorting to such childish behavior. He must feel incredibly frustrated that no amount of money is able to make him more competitive within the AI space.

I think it would have been better to have just brought Ashok Elluswamy over and placed him in charge of a group and then tried to just keep the researchers on rather than firing them. It is hard to get anything done if you do not have the talent already onboard.

How come all the departed researchers are Chinese nationals?

lol! no surer sign of a junior/naive/ignorant developer or manager than the sentiment "okay, well, let's start from scratch and do it right this time."

big projects generate cruft. there are ways to minimize it, but as you go along there will always be some stuff that doesn't quite mesh with whatever else you've got going on. if you insist on ironing out every single wrinkle (admirable!) you'll never actually deliver a result.

I'm not saying this will fail. green field projects can certainly be a godsend when they produce something better than what they attempt to replace. but they are always a sign of failure. of not being able to work your way out of the mess you made with the first attempt. so that just begs the question: what are you going to do when this attempt gets hard to work with? going to give up and start over again - do it right that time? or...?

There should be a social theory study of jerks and people who continue to work for them. I’d imagine there’s a sort of logarithmic drop off. It takes time for the knowledge of terrible bosses to work through the labor force, but tech is a small community (of hopefully smart people) and people start to catch on.

A story I heard from a Tesla employee was that it’s impossible to hire because musk is mercurial and every time a hire makes it through the pipeline a hiring freeze cancels all offers. Another story that employees were told to avoid his desk because he randomly fires people. Another that he regularly cancels bonuses because he’s feeling petulant. I heard another story that he threw up Nazi salutes on an internationally televised event. No wait, that one I watched happen.

After a while smart people will simply decline to work for him no matter the compensation. Not everyone, but enough to start to matter.

Unfortunate. The Grok team built a phenomenal model. I use it all the time and it very often out performs GPT and Claude, on coding and STEM research related tasks. I was part of the beta for a while Grok 4.2 Beta with multi-agents and it was just amazingly good.

People aren't using it for reasons other than its capabilities. I mean, I don't think my boss would approve a paid Grok subscription for example.

@grok is this real?

@grok fire the bottom 50% engineers from x.ai ranked by number of commits per day

@grok generate a hypothetical picture of an Elon who is not under the influence of large amounts of Ketamine

I honestly don't know what to expect from Elon these days. But it's rarely good news.

Musk sounds like such a nightmare to work for. I legitimately don't understand why anyone would put up with him. What's the appeal?

Did everyone forget that xAI was just a way for Musk to weasle his way out of debt from his Twitter purchase?

It does not surprise me. The free Grok got worse since 4.0, they increasingly save money by not responding at all or only allowing one answer. Grok now defends the administration and billionaires.

The company seems to burn money like crazy. Everyone knows that "AI in space" and the downgrade to a moon trip after claiming for 15 years that Mars is just around the corner are marketing.

All AIs are toys and the coding promises are just a lie to string along investors. Unfortunately many of these are senile Star Trek watchers who buy into everything.

I do use Grok as a chatbot sometimes. Very good for sourcing X and general web search. Not as "prude" as the others too.

Maybe SpaceX will buy xAI and Tesla to hide the systemic problems at his companies into the warm embrace of the one legit useful company.

The pleasure people get at seeing Elon's companies failing is astounding.

This might explain why Grok went unavailable to non-subscribers at X the other day.

MechaHitler and undressing aside, Grok is the best integration of AI into an existing website ever attempted.

The Takeover by SpaceX was obviously a Bailout. And now they pressure NASDAQ to change the rules so they can dump their junk into the index funds.

Wow, bit weird that Musk, who must have known about how badly xAI was doing, spent so much of his investors money buying out xAI.

What an enormous blunder.

Wait, what does this imply for Cursor? I DGAF about xAI and will never use their Grok, but I did like Cursor more than the alternatives (even if I'm just running opus 4.6 most of the time).

But now he is poaching the two heads of engineering of a company that's trying to move very quickly, how is that going to affect their speed and success?

Ali think all of Elon’s companies have an Elon problem: he’s so polarizing he’s limiting the talent pool to choose from. I’m here for it because I don’t care for the man and want him to fail… but yeah that seems clear to me that his polarizing antics are costing his companies.

Where is the grok coding cli?

Will this be an indictment on the insane work hours I've heard the xai team pulls?

I feel like even just a couple years ago it would have been shocking to see an article involving Musk have this kind of spin. Like you'd never see a line like this:

> The name is a “funny” reference to Microsoft, the billionaire added.

in something from 2023 or earlier.

xAI showed me that it’s really still OAI and Anthropic (which is basically the OG devs). No matter how much money you throw at the problem, the entire space is still in the hands of a few.

im not surprised, grok definitely falls behind as both a coding agent and a research tool.

claude codes the best, gpt is the best research tool, and grok is really only great at videos. which isn't a huge loss, but videos don't have the same functional capacity as academic topics and coding

It's surprising that AI coding agents have network effects but it's true. Think about it from first principles & you'll realize that the bottleneck is how many people are using it to write real code & providing both implicit (compiler errors, test failures, crash logs, etc) & direct ("did not properly follow instructions", "deleted main databases", "didn't properly use a tool", etc) feedback. No one is using xAI for serious software engineering so that leaves OpenAI, Anthropic, & Google w/ enough scale to benefit from network effects. No one has real AI but what they do have is the appearance of intelligence from crowdsourced feedback & filtering. This means companies that are already in the lead will continue to stay there & xAI started way too late so they will continue to lose in every domain that actually matters & benefits from network effects.

Not even Elon believes that Cursor is worth $50B or even $29B.

Their goal of moving compute to space combined with their capacity to launch tons of payload will make this look like a tiny blip.

I couldn't find a working archive link for the ft.com article - anyone?

Since it's the original source I've left it up, but added other URLs to the toptext.

[stub for generic-indignant tangents - not what this site is for - please see https://news.ycombinator.com/newsguidelines.html]

dang wrote:

> You may not owe you-know-whom better, but you owe this community better if you're participating in it.

This is like telling a country that’s being invaded that they can only respond with strongly worded letters when their enemy is dropping tactical nukes on them.

But hey, Paul Graham and cronies benefit from the status quo as much as any other billionaire, so let’s not rock the boat, right?

The word “complicit” comes to mind.

The grok button on twitter is pretty awesome. Instantly summarize / explain any tweet, even memes, including replies. Ask follow up questions. Not sure many people know it's there.

Also grok in the Tesla is fun, get answers to questions without looking at a phone. I once had it search up a blog post and read it out to me while driving. The NSFW mode is pretty...disgusting so I leave that off.

I hope they find a way with Optimus or something. FSD is incredible. More competition is a good thing.

Yes 11 up and everyone why free insult on a model that top adoption. Aligned with your personal view is not ahead of the curve, it's just personal.