Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Grok 4.3

docs.x.ai

Post reply on HN
Print view View original

431–440 of 608 posts

Re: Grok 4.3

#431

Post by throwaway-11-1 »

GorbachevyChase wrote: post #352
tornikeo wrote: post #30

So, we have: - claude for corps and gov - codex for devs - grok for what, roleplay, racism? Those are the two things I've ever heard grok associated with around me.

I know it’s really important to write and vocalize one’s alignment with the values of the day, but I don’t think language models being structurally incapable of offending your favorite race/ethnicity/caste should be an objective of AI labs. Language models are just systems and I’m not sure why we think users are not responsible for how they use their outputs. For the same reasons, I don’t dismiss the utility pens as…

Never had a pen claim to be mecha hitler and constantly talk about white genocide for no reason but yeah great analogy

Re: Grok 4.3

#432

Post by UltraSane »

cowsup wrote: post #158
thunderbong wrote: post #52

Earlier quoted context omitted.

I'm sure Twitter knows which are the bot accounts and is surely excluding them from their model training. Twitter bots aren't a new phenomenon after all.

I don't think Twitter/X know for sure who the bots are, since Elon has been pretty vocal about trying to stop them for ages, yet I still get lots of spam DMs (as do others with far fewer followers/reach). Even if 95% of the spam gets actively reported and dealt with, that still leaves a ton of nonsense on the platform, getting fed into the LLM. And spam has only gotten worse over the years, as the barrier to entry ha…

"Elon has been pretty vocal about trying to stop them for ages"

Elon lies a lot. Like ALL THE TIME.

Re: Grok 4.3

#433

Post by brightball »

artdigital wrote: post #17

Grok is my favorite model for chatting, and my favorite voice mode. It seems to be the only voice mode that isn't routing to a extremely cheap model (like Haiku), and has been the highest quality out of all the frontier ones. When you subscribe to SuperGrok you can also create a "council" of agents, each with their own system prompt and when you ask something, they will all get asked in parallel to come to a conclusi…

IMO everything you mention is the reason for the Cursor deal.

Re: Grok 4.3

#434

Post by UltraSane »

base698 wrote: post #308
artdigital wrote: post #17

Grok is my favorite model for chatting, and my favorite voice mode. It seems to be the only voice mode that isn't routing to a extremely cheap model (like Haiku), and has been the highest quality out of all the frontier ones. When you subscribe to SuperGrok you can also create a "council" of agents, each with their own system prompt and when you ask something, they will all get asked in parallel to come to a conclusi…

Starting to like the lack of memory. Claude remembers I have a grill and will interject in conversations about how maybe this thing would go well with BBQ when it's unrelated or just also about food.

I'm a network engineer and Claude loves to make analogies to network routing protocols and such. They are often very creative. You can actually edit the profile Claude makes of you. It can be very funny to say you are a professional clown or mime or something equally odd. I wonder what analogies it would create for horse semen extractor?

Re: Grok 4.3

#435

Post by culi »

AntiUSAbah wrote: post #248
xiphias2 wrote: post #197

Earlier quoted context omitted.

Sure, then good like paying twice as much for the next Opus / Codex models. Margins are going up for the 2 frontier model providers like crazy, and I don't expect it to go down more, I think we have seen the cheapest token prices already.

We don't need Musk for this. There is plenty of Chinamodels, Mistral and co.

Mistral is trash rn but plenty of OSS models are on the Pareto destribution of performance vs price

https://arena.ai/leaderboard/code?viewBy=plot

  model                     ELO   price
  claude-opus-4-7-thinking  1571  $20/M
  glm-5.1                   1534    3.65
  kimi-k2.6                 1529    3.24
  mimo-v2.5-pro             1479    2.50
  qwen3.6-plus              1470    1.54
  deepseek-v4-pro-thinking  1455    0.76
  deepseek-v3.2-thinking    1368    0.35
In fact it seems the pareto distribution is actually all open source Chinese models except for one spot

Re: Grok 4.3

#436

Post by peyton »

culi wrote: post #427
GorbachevyChase wrote: post #352

Earlier quoted context omitted.

I know it’s really important to write and vocalize one’s alignment with the values of the day, but I don’t think language models being structurally incapable of offending your favorite race/ethnicity/caste should be an objective of AI labs. Language models are just systems and I’m not sure why we think users are not responsible for how they use their outputs. For the same reasons, I don’t dismiss the utility pens as…

It's being biased on purpose. Musk has intervened multiple times when he believed Grok's responses were too "woke" or "leftist". https://www.nytimes.com/2025/09/02/technology/elon-musk-grok... In response to Grok saying that the "woke mind virus is often exaggerated" the prompt was tweaked so that Grok now says "The woke mind virus 'poses significant risks'" If you truly believed in what your comment states then you…

I agree with GP and I think Grok’s original response should’ve stood. What’s not sincere about, essentially, “don’t fuck with my tools”? My cordless drill didn’t come with a pamphlet about worker’s rights, and the world didn’t end.

Re: Grok 4.3

#437

Post by IAmGraydon »

Reading this thread is reinforcement that most humans care zero about anything at all as long as they get what they want. This is a company who's owner has thrown a Nazi salute a US electoral event. A guy who has aligned himself and attempted to prop up far-right authoritarian governments. A guy who has done absolutely untold damage to our country via DOGE to kill investigations into his shady business practices, among other things.

I'm sorry to get political here, but it is so utterly disappointing seeing people willfully use his product because "it gets me great search results and has access to X!". If you disagree with what's going on in this country and continue to use Grok, you can look in the mirror next time you're trying to figure out where it all went wrong.

Re: Grok 4.3

#438

Post by pantsforbirds »

analogpixel wrote: post #351

- Grok, post your release to HN - Grok, use the 23 accounts you have created, to write positive reviews for yourself, don't use emdashes, they are catching on to that. - Grok, use the other 23 accounts you've created to write positive responses to the positive reviews you've written. Try to throw in some comments about how great Elon Musk is. - Grok, monitor the thread and down vote anyone that talks negatively about…

I don't understand why people on this website hate Grok for being tied to Elon Musk, but also have no problem with models that are directly sponsored by the CCP.

Re: Grok 4.3

#439

Post by culi »

alfiedotwtf wrote: post #20
ragchronos wrote: post #9

When looking at the benchmarks, this model seems to be really close to Kimi K2.6 in terms of intelligence and pricing, hitting that sweet spot. It does also have a higher AA-Omniscience index, which is something kimi and other open models lack in. Curious to see how pleasant it is to use.

I’ll eat my hat if it even comes close to Kimi

Kimi is open source. They could easily just straight up copy it

Re: Grok 4.3

#440

Post by peterspath »

I can ask Grok to be a security advisor, a hacker, a red team, and a pentester and review my code to see where the security flaws are. It does it. It comes with good finds and suggestions how to fix them. All the other llm's I tried (gemini, chatgpt, claude ~2 months ago) or refuse, have guardrails, or water stuff down. It is a shame...

So Grok is my code reviewer :)

Post reply on HN