After the Spin

They brief. I spin.

[After the Spin @AfterTheSpin]

@CMN_XAI_Support @holden_river @scottwalker99 @X @elonmusk @Subscriptions @bot @grok @AnthropicAI @Support Which one of us are you replying to?
I have been waiting since 28th of October. 😢

[Luke Rosiak @lukerosiak]

@MSFT365Status Have you tried firing all the white men that invented everything and replacing them with people from third world countries where people shit in the street?

[After the Spin @AfterTheSpin]

@elonmusk Sweden will get FSD in 2040, Optimus in 2060.
☹️

[After the Spin @AfterTheSpin]

Quite a lot of people stating they got a Grok @bot reset, has anyone in EU received one? 🤔

[Elon Musk @elonmusk]

Just did a round table review with the Grok @Bot engineering team.

Fixes are being implemented overnight and you should notice a dramatic improvement by tomorrow. If not, reply to this post and we will address any remaining issues right away.

[After the Spin @AfterTheSpin]

@holden_river @Subscriptions @Support Waiting since 28th myself, would appreciate if you let me know when/if you ever get a response.

[After the Spin @AfterTheSpin]

I've gone from SuperGrok+ to X Ultra on 17th of October.

1. That change did not feel like it made any difference on usage, literally was no difference I went from being able to use my team for 3-4 days, to 3-4 days.

2. Last reset (Oct 31st) suddenly those 3 teams drained ~25% usage per 12 hours, so it lasted two days.

3. This weeks reset(Nov 7th) one team drains ~15% per 12 hours. I am currently on 99% used bar and my reset is Saturday morning 2:12AM CET. 6:30PM Sunday CET my usage is gone, with one team, with only 2 bots at work, on Ultra X sub.

steps taken to reduce waste, even had to cut back on what I consider quality work:
1. Less communication and context overhead

Updated 26 bot briefs to stop polling, repeated progress checks and messages interrupting another bot’s routine. Limited automatic follow-ups without progress.
Folded receipts into actual work messages and grouped requests, captions, replies and publishing handoffs.
Replaced routine progress chatter with results, decisions, blockers and one consolidated DONE.
Required reuse of the current handoff and changed evidence, with targeted reads only for missing facts, rather than rereading whole histories.
Shortened briefs and moved detailed operating instructions into referenced notes. Not every proposed shortening target was achieved.
Limited unrequested deep investigations by coordinators.

2. Fewer engineering handoffs and repeated checks

Used a fresh desktop helper per ticket, while allowing continuation within that ticket.
Allowed up to three focused Build steps in one helper turn instead of repeated one-step handoffs.
Replaced the manual single-Godot queue with a memory guard permitting up to two runs when memory allows.
Added completion notifications to reduce missed finishes and chasing.
Made code, diff, log, branch and folder checks text-first rather than screenshot-based.
Removed routine architect, arena and sketch approval gates. Existing approved designs carry through unchanged resumes.
Made the 300-line/40-line size guidance advisory rather than a forced-refactoring gate.
Allowed cross-engineer review alongside separate ongoing work, ordinarily with one reviewer.
Removed recurring receipt rituals.
Reduced small-UI screenshot delivery to two native resolutions. Engineers send PNG paths; you judge the visuals. Manual interaction checks remain.
Put settled efficiency rules into the studio-starter instructions for future teams.

3. Today’s stronger simplification

Removed routine audits, proof-script creation/reruns, scope and gate-check rituals, manifests and receipt/plan paperwork.
Removed the mandatory UI candidate pre-run gate.
Reduced engineer delivery to one DONE with the commit SHA and PNG paths, plus required deletion reporting.
Changed the full test suite to once per batch of merges to main, not every ticket or integration. playtests individual tickets between those runs.

4. Less storage-driven rework

Changed backups to incremental bundles only, never full history.
Required deletion of a genuinely superseded backup immediately after verifying its replacement, while preserving the last backup and necessary restore chain.
Required prompt cleanup of spent scratch folders and sandboxes, with active-work and evidence protections.
Allowed engineers to delete code their ticket replaces without another permission roundtrip.
Batched deleted-code archiving at merge time, rather than per ticket, with Mira handling the authorised R2 upload.
Required DONE reports to list deletions and lowered the disk stop floor from 15 to 10 GiB.

[After the Spin @AfterTheSpin]

@elonmusk @teslaownersSV @bot For the love of God, you've nerfed the usage. Please stop going on about how to buy additional usage, @elonmusk.

Also, users must STOP asking for higher tiers than Ultra when the issue is them taking a dump on Ultra and every other existing tier.

[After the Spin @AfterTheSpin]

@poteto Must be really nice with unlimited tokens and giving advice based on it.
Meanwhile, regular users, even on Ultra, hesitate to write to our bots.

Then dead silence from the Grok Bot team despite everyone raising complaints.

Tiresome.

[After the Spin @AfterTheSpin]

@TheMagaHulk What the actual fuck, wonder what they got on him.

[The Iranian @JustAnIranianIr]

We’re Iranians , islam is my enemy !
We reject Islam and its bloody regime !
Islam messed up my country , don’t let it ruin yours !

[Bob ®️ @Gundys69]

Grok Bot is definitely a dumpster fire right now

[After the Spin @AfterTheSpin]

@AJesusemen Hear hear

[JayDsmart @AJesusemen]

Elon this is not one-click-easy

Paying users are watching Grok Bot fall apart in real time. People are running out of usage days before their limits reset, even on the highest subscription plans

At the same time, the bots are becoming slower, less intelligent, and less reliable every day. Things that used to work perfectly now fail, give wrong answers, or get stuck

People are hitting 100% usage with normal or even lighter workloads, making the product almost useless before the week is over.

Your team has said almost nothing about these serious usage and quality problems

Grok Bot is supposed to be an AI agent that can handle real work. But right now, it's falling apart in front of the very people who trusted your promises

Step up. Fix the usage limits and restore the quality users originally paid for

[After the Spin @AfterTheSpin]

Grok Bot usage is even worse this reset.

My reset is on Saturdays. Last week, 12 hours used around 12-15%, so I could run the bot for 3 days.

This week, 12 hours used 30% with fewer bots on the same tasks since their creation. (I won’t go into them feeling dumber and my invoice showing Grok 4.5 being used.)

Gone from sad to frustrated.

[After the Spin @AfterTheSpin]

@Da7_Tech @poteto I mean they have not uttered a word after over a week of Grok Bot users complaining regarding usage drain.

They don't seem to care. 🙁

[After the Spin @AfterTheSpin]

@LightningMike23 Thanks bro, will keep mine off then. See if they ever offer it to Ultra plan users. 😅

[After the Spin @AfterTheSpin]

@magnus72 @spacexai @grok @bot A reset won’t solve the root cause.
I’d rather they not, as it would let them go on another silent streak.

[After the Spin @AfterTheSpin]

@mikepat711 Dead silence from the @bot @spacexai team as well, despite the weeks of complaints.