Marcos
5.9K posts
By far one of my favorite features of Grok Build is how it handles conversation compacting.
With Claude Code, compacting can take ages. With Grok Build it tells you it compacted, then instantly keeps building.
That alone is a huge part of why this thing cooks so fast!
English
@nikitabier @dieworkwear it is a cat and mouse game, the are no rules nor they stay static, it is a endless mission. you cannot state "there is no" if "they" successfully accomplish their game.
English
When I first started at X, this is one of the things I did a deep dive on because I felt it was critical to the integrity of the experienced.
I tasked our Threat Disruption team to investigate a number of trends that seemed artificial.
The findings:
We could not find meaningful examples of foreign interference in US policy discussions, except people gaming rev share in developing countries. This is what motivated the release of the Country of Origin feature and significant changes to the rev share algorithm.
The most deranged & divisive replies generally were from residential IPs in the United States—with no signs of using a VPN.
Ultimately, X is a reflection of the internet. And that means you will see the full spectrum of human thought. And sometimes the most outrageous takes will catch fire.
Having said all of this, there can still be cases of narratives being boosted but the origin of the initial post is almost always domestic and we have hardened our systems in the last 3 months to prevent this.
English
@WesEklund @grok note that the promo is not that you pay 99 per month on a monthly basis, you have to pay the 3 months together! that might be the confusing part
English
How do I leverage the SuperGrok Heavy promotion?
Clicking on "Claim Offer" says "Go to the Play Store on mobile", but the Grok app only lets me do the normal pricing 🤔
This is my only blocker for @grok SuperGrok Heavy
English
@tetsuoai I don't have grok 4.5 yet (Europe) but /goal has been a pain in the ass because of is skeptic reviewer loops backing off too frequently because can't resolve disputes. Los of tokens are wasted on back and forth disagreements about goal being implemented or not.
English
I only use Grok 4.5 now.
The /goal command in Grok Build is changing how I actually ship things. It is very fast.
It plans the architecture, writes the code, spins up sub-agents, tests everything, fixes its mistakes, and runs autonomously until the goal is complete and verified.
Ask Grok to write documentation for a full feature in a plan.md file. Then, have it create a ‘goal’ prompt for that plan and send it back using /goal
English
@yunta_tsai @ShipInDistress @patnrv are talking about the hidden layers? this phrase "If the model hides the thinking traces without telling you, then you will feel the jump." make it sound like it is deliberate as per the LLM choice, which i think is rather by design
English
@ShipInDistress @patnrv There is a lot of continuous hard work you don’t see.
Many of my ideas come from shower thoughts, but I’ve been banging my head over them for years in the background.
English
I would like to offer a counterargument that LLMs (or maybe AIs) cannot jump.
Before AlphaGo, the AI field had the same argument for Go: there are 2.08 × 10^170 possibilities, nothing fits in the computer, and there is no way AI could possibly predict the outcome of the next 50-60 moves.
It turned out most moves do not lead to a win. Combined with clever use of Monte Carlo Tree Search, the sampling becomes quite manageable. The same can be said for physics, where equations are just another form of compression.
Einstein did not start with relativity. That was not his first paper. He spent years understanding the properties of light before concluding that the speed of light is constant across the universe, which unlocked his discovery of relativity. During his thought process, he also interacted with other physicists (e.g., sub-agents) to enrich his thinking.
Currently we have not run an agent for years of compute. The sessions are often fragmented and disoriented, so every new session is almost a fragmented memory of the past, but it may not be for long.
Han Xiao ✈️ ICML 2026@hxiao
interesting position paper throwing cold water on autoresearch/ai scientist: LLMs can't jump. The thought experiment is this: Take an LLM with a 1905 knowledge cutoff. Feed it every paper, every dataset, every equation of that era. Could it invent general relativity? No. Discovery isn't one thing. It's three. You can induce — generalize from data, which lands you at Newton plus some epicycles to explain Mercury's weird orbit. You can deduce — derive rigorously from axioms you already have, which never gives you new axioms. Or you can jump — invent the frame itself, decide that spacetime curves. That third move is the one that matters, and it's exactly the one induction and deduction can't reach. Penrose put it as three worlds: Physical, Mental, Platonic. Data flows from the world into a mind fine. But the new law has to be discovered into the Platonic world first — and that step is the jump. LLMs are induction machines running over what already exists. Structurally, they don't take it. I think it’s a warning to AI scientists/autoresearch against collapsing two very different things into one word. Hill-climbing: LLMs are already superhuman here, and autoresearch in this sense is real and moving fast. Abduction/leap/jump: a new frame that reorganizes the field, that is a different act entirely, and nothing about scaling induction suggests you get there. Most of what Autoresearch ships today will be spectacular hill-climbing. The jump is still ours for now.
English
Also: "Hebrew University of Jerusalem owns the rights to Albert Einstein's name and image, having secured a trademark on the Einstein name in 2003. They actively manage licensing agreements and have earned significant revenue from the use of his likeness in various commercial contexts."
English
Einstein is way over rated and most of his work is actually based on others, (like most). "Albert Einstein presented the theories of special relativity and general relativity in publications that either contained no formal references to previous literature, or referred only to a small number of his predecessors for fundamental results on which he based his theories, most notably to the work of Henri Poincaré and Hendrik Lorentz for special relativity, and to the work of David Hilbert, Carl F. Gauss, Bernhard Riemann, and Ernst Mach for general relativity" + "In 1905, three articles appeared in the 'Annalen der Physik', which began three very important branches of 20th century physics. Those were the theory of Brownian motion, the photon theory of light, and the theory of relativity. The author of these articles – an unknown person at that time, was a bureaucrat at the Patent Office in Bern, Einstein-Marity (Marity the maiden name of his wife, which by Swiss custom is added to the husband's family name)" "Einstein Works Out Details of His 1919 Divorce from Mileva Marić". Shapell Manuscript Collection. Shapell Manuscript Foundation"
English
@namitooto i used to grab the korg volca, the bike, drive and play =) around. btw nice composition too. are you feeding the composition from the mic and live sampling or is it just for ambience rec?
English
江の島水族館の前にグルボ用の台があったから音遊び♪ココは車の動きがあって良いなぁ♪ちょっとづつModelCyclesに慣れていっている気はする♪
日本語
GAMERS!! Name 5 Commodore 64 games you think everyone should play at least once.
I'll start:
1) Elite
2) Maniac Mansion
3) Impossible Mission
4) The Last Ninja
5) Sid Meier's Pirates!
English
@EV_Trapper grok might have autoloaded your agent files meant for other platforms and using them all, happened to me.
English
Yeah, FABLE-5 is considerably better in every metric except for compute efficiency than Chatgpt 5.6 SOL.
Grok 4.5 is significantly faster than both, but is honestly around OPUS 4.6. There is something really wrong with the usage for GROK 4.5 though, less than 10 prompts & it used up an entire week's worth of usage!
Whole Mars Catalog@wholemars
because fable is eating his lunch and grok is eating his breakfast
English
@ZachWarunek @BrianEastwoodx editor+imagine for the missing hand gestures= win (should be 4 premium+ only)
English
@BrianEastwoodx Good video. FYI you might wanna put the cc a bit higher so it’s above the bar. We should probably give some indication in this editor ui if we don’t already
English
I am not profoundly unhappy with myself!
Please seek help.
English
@LouiseGiam @Laz252 @michaelgofman @elonmusk @SpaceXAI the billing is correct, you 300 once, you get 3 months.
English
@FlyingLow666 un poco de historia: los ingleses ya estaban hacia mas de 150 años en la isla, argentina ni tenia la patagonia
Español
@PeterDaveHello @SpaceXAI it would be nice that they develop an internal tool that values feedback into score and regard the useful and great ones credits @Engineering @Support also, the users that are opt-in should have more allowance that the ones that are not
English
Been putting Grok 4.5 through its paces these past few days, and it’s genuinely impressive, especially for the price.
Would love to see @SpaceXAI give subscribers a little more room each week. That extra headroom would let more people experiment with Grok Build and put Grok 4.5 to work on all kinds of real-world tasks. Feels like a smart way to win over even more users.
Already hit my weekly limit, though. 😅
English
Tell me you mix music without telling me that you mix music
English