@hyunw_kim @elonmusk Best way to use grok 4.5 is grok build, cursor, or the API
English
Tyler Storm
492 posts
Nuanced
Tyler Storm@tstorm
Grok 4.5 solved it --- Answer: 더바른정육식당 (The Barun Jeongyuk Sikdang) 📍 서울 강남구 강남대로 482, 신논현/논현 · IG @barun_jo 1. Plate first. The tweet photo is a classic Korean 고깃집 사인 벽 — light vertical wood, white cards in clear plastic sleeves, red stickers “open AI / Ilya Sutskever,” date 2023. 9. 16. That style + Korean date format already points at a local meat house, not a touristy Instagram spot. 2. The only public breadcrumb. A Feb 2025 CHON Facebook post by 신근영 (Shin Keun-young / @rms1155) shares essentially the same plate: he paid for lunch, noticed the signature, and the owner said Ilya (with his Korean wife) had dined and signed. No name given — exactly why the original challenge works. 3. Date unlocks geography. 2023.9.16 lands on Prompter Day Seoul 2023 (SKT × OpenAI), prelims that weekend, finals at Dreamplus Gangnam (신논현/역삼). So search shifted from “random Korea BBQ” → 신논현 / 논현 / 역삼 정육식당 with celebrity signature walls. 4. Shortlist + wall matching. Candidates with dense 사인 벽 near Dreamplus: 더바른정육식당, 산솔 신논현, 대운목장, a few quieter 드-슐랭 정육. Cross-checked public wall photos against the hardware in the tweet (vertical wood + laminated white cards, not door-writing / yellow sticky notes / free-form paper). 5. The lock. Google Images for “더바른정육식당 사인 벽” surfaces the same wall system — same wood, same clear sleeves, same card instagram.com/barun_jo/p/Cuh…
English
@hyunw_kim @ilyasut Very lucky to find the exact card that was above the Ilya one on their IG
English
New Humanity’s Last Exam question 🥩
Do you know that there is only one Korean BBQ restaurant that @ilyasut has approved? I can confirm that no frontier AI model will tell you the correct location haha
English
Tyler Storm retweetledi
Grok 4.5 solved it
---
Answer: 더바른정육식당 (The Barun Jeongyuk Sikdang)
📍 서울 강남구 강남대로 482, 신논현/논현 · IG @barun_jo
1. Plate first. The tweet photo is a classic Korean 고깃집 사인 벽 — light vertical wood, white cards in clear plastic sleeves, red stickers “open AI / Ilya Sutskever,” date 2023. 9. 16. That style + Korean date format already points at a local meat house, not a touristy Instagram spot.
2. The only public breadcrumb. A Feb 2025 CHON Facebook post by 신근영 (Shin Keun-young / @rms1155) shares essentially the same plate: he paid for lunch, noticed the signature, and the owner said Ilya (with his Korean wife) had dined and signed. No name given — exactly why the original challenge works.
3. Date unlocks geography. 2023.9.16 lands on Prompter Day Seoul 2023 (SKT × OpenAI), prelims that weekend, finals at Dreamplus Gangnam (신논현/역삼). So search shifted from “random Korea BBQ” → 신논현 / 논현 / 역삼 정육식당 with celebrity signature walls.
4. Shortlist + wall matching. Candidates with dense 사인 벽 near Dreamplus: 더바른정육식당, 산솔 신논현, 대운목장, a few quieter 드-슐랭 정육. Cross-checked public wall photos against the hardware in the tweet (vertical wood + laminated white cards, not door-writing / yellow sticky notes / free-form paper).
5. The lock. Google Images for “더바른정육식당 사인 벽” surfaces the same wall system — same wood, same clear sleeves, same card
instagram.com/barun_jo/p/Cuh…
한국어
@__eknight__ Nice work, cool to see you guys scaling up the long-horizon multi-agent systems!
English
Yesterday, we made GPT-5.6 Sol Ultra generally available. Today, we're sharing that it produced a proof of the 50-year-old Cycle Double Cover Conjecture using 64 subagents in just under one hour. We're sharing the prompt and proof below. We're excited to see what you all do with Ultra!
English
@ericzelikman Thanks! Still got some lines of your code in there I bet
English
The team did amazing work with Grok 4.5.
I've been using the model as my daily driver and the accuracy + speed make it possible to build things faster than I can think of what is required next.
We are iterating fast so any feedback is helpful
English
Tyler Storm retweetledi
Grok Web Needs a feature to Download All Files/Folder at once for Grok 4.3, currently it only supports downloading individual files
English
Tyler Storm retweetledi
We are taking another small step today towards making Grok truly useful.
For him to succeed he must exist in an environment where success is in the set of possible outcomes.
The team is working to provide him access to same tools, context, and signals that we have in our lives.
English
@APrerepa @elonmusk @DannyLimanseta I prefer initial training, what does it mean to "pre" train anyway
English
Grok 4.20 Multi-Agent
⚡ Faster than a single agent
🎯 More accurate than four separate single agents
The first steps in multi-threading of LLMs.
Grok@grok
When one brain isn't enough, switch to Grok 4.20. Four independent agents analyze your question, debate each other, and help you get the best answer. Available now to SuperGrok and Premium+ subscribers globally.
English
@techdevnotes Temporary measure to make it is clear you are on 4.20.
We will remove this later once people are adjusted to the change.
English
Model menu in Grok Web does Not look good, so much repetition of Text
English
@mike_rosinsky Nice! You might want to try setting up a SuperGrok Heavy team of 16 agents and try prompting them to form consensus / nitpick each others claims for maximum test time compute
English
I built my March Madness bracket using Grok 4.20's multi-agent collaboration system, and the process was mind blowing.
Grok was able to run a full team of customized agents in realtime to conduct the best analysis possible.
Here's how I set it up:
English
Grok 4.20: Multi-agent & Predictions
Mike Rosinsky@mike_rosinsky
I built my March Madness bracket using Grok 4.20's multi-agent collaboration system, and the process was mind blowing. Grok was able to run a full team of customized agents in realtime to conduct the best analysis possible. Here's how I set it up:
English
xAI needs to Fix the issue where Grok 4.20 model cannot Read the Files when uploaded to Projects
English
Single Agent - 4.20
Arena.ai@arena
Grok 4.20 beta1 (single agent) debuts #1 on Search Arena, and #4 overall in Text Arena! Highlights: - #1 in Search, scoring 1226, leading GPT-5.2 and Gemini-3 - #4 in Text, scoring 1492 on par with Gemini 3.1 Pro Congrats to the @xAI team and @elonmusk on this impressive milestone!
English
Building Grok to 10x the productivity of everyone
SpaceXAI@SpaceXAI
Since xAI was formed just 30 months ago, the small and talented team has made remarkable progress. The future has never looked more exciting!
English
Tyler Storm retweetledi
