@Kenny_GTM I believe you could argue the “AI agencies”that was the hype in late 2024 and 2025 is now GTM engineers in late 2026.
I’d even say that they are relatively the same thing but just different names
@grokbot@cursor_ai quick q for the builders 👀
Just got back into coding after some time away, and "Projects" wasn't out yet when I left.
What I've noticed since coming back:
• Lots of people are using Grok Bot to manage their Cursor agents, but it doesn't feel like the coding tool Cursor built
• My repos live on @GoEntire, and I'm running into issues with cloud agents there
How are you all setting up your workflow? Would love some inspo 🙏
@ahamam101@GrokBot@cursor_ai Hey Ahmed!
Currently building a bunch of side projects that I’ll eventually open source.
I have a passion for Anime/manga/books so currently making a better version of My Anime List. It needs an update!!
Been building with @GoEntire for the past hour and hit one roadblock that confused me at first.
On GitHub it's simple: one person, a bunch of repos.
Entire goes one level deeper: 📚 Your profile is the library 🗄️ Each project is a bookshelf 📖 The repos are the books on that shelf 🧭 And each shelf holds all the trails for its books
Once I pictured it like that, it clicked. Anyone else have to rethink how they organize stuff?
I tidied up your bookshelf analogy so each level matches one thing. If the trails in Entire belong to each repo instead of each project, tell me and I'll fix that line.
I have a major result to share.
I’ve found conclusive forensic evidence that the top 4 Chinese AI labs—DeepSeek, Moonshot, MiniMax and Ant Group—have built their newest models on NVIDIA’s most advanced Blackwell chips, which are barred from export to China.
This sounds insane, but I discovered that almost every AI model carries a hidden signature of the GPU that trained it. When a model is first set up, the starting values of its weights are laid out in a pattern that depends on the exact GPU. Each combination of GPU and seed number produces a unique pattern—a sort of fingerprint of the hardware used to create the model. Traces of that pattern survive training into the published model.
Again, this evidence isn’t based on any insider information or leaking. It’s based on a forensic analysis of the model weights released by the companies themselves.
I searched more than six trillion combinations of chip layout and seed number and matched them to open-weights models. In doing so, I found conclusive evidence that six Chinese AI models—including DeepSeek-V4-Pro, Kimi K3, and MiniMax-M3—were initialized on NVIDIA Blackwell systems. Four of the models match Grace Blackwell, the GB200 and GB300 parts in NVIDIA’s flagship liquid-cooled rack-scale systems.
I then verified these results on real hardware. Running unmodified PyTorch with the seeds I’d recovered, a Blackwell B300 reproduced DeepSeek-V4-Flash’s untouched starting weights bit for bit, and a rented GB200 reproduced Kimi K3’s to a correlation of 0.999995. A Hopper H100 reproduced none of DeepSeek-V4-Flash’s, and changing the seed by one broke the match. (DeepSeek’s seed is 953049583 by the way.) The method also was verified to match 24 known open-weights models to the hardware that developers said they used.
It is exceedingly unlikely that these results are a coincidence. Across six trillion combinations, the best accidental match should reach only about 8 sigma (i.e., eight standard deviations above what random chance would produce). The real matches range from 49 to 1,385 sigma, and even the weakest single test hits 20.
These were data-center-scale training runs. With some analysis of release schedules and model details, I estimate that the largest, Kimi K3 and DeepSeek-V4-Pro, used 2,000 to 3,000 Blackwell GPUs each. Clusters of this size cost on the order of $150 to $200 million each to buy, and across these labs alone would add up to more than half a billion dollars of export-controlled GPUs, not counting experimental compute or otherwise. It’s quite likely that we see multi-billion dollar expenditure on export-controlled GPUs.
I am highly confident in this result, and it has been independently verified by senior researchers at OpenAI and Anthropic. You can read the full report at https://t.co/YOtzND0Jqk.
There is much more I’ve identified about these models through other forms of checkpoint forensics like this, and will have more to share soon.
Looks like your Mac CI runner just became optional.
e2e runs iOS and Android tests on Expo Simulators from Linux, on every pull request.
There are a lot of reasons to love this. Tough to narrow it down...more confidence in every PR is probably the best thing.
ever had your PC do something that looked like a hack? mine did 3 times in one day lol
• it asked for a BitLocker recovery key out of nowhere. turns out Windows was installing the new Secure Boot certificate update, the laptop crashed partway through, and BitLocker locked the drive to be safe
• a PowerShell window popped up every time I logged in. looked like malware. it was 𝘮𝘺 𝘰𝘸𝘯 𝘴𝘤𝘳𝘪𝘱𝘵 for mounting a dev drive, failing because it didn't have admin rights
• ran a full check after. no malware, no weird accounts, nothing strange at startup or connecting out
what to do before you panic:
• 𝘀𝗮𝘃𝗲 𝘆𝗼𝘂𝗿 𝗕𝗶𝘁𝗟𝗼𝗰𝗸𝗲𝗿 𝗸𝗲𝘆 in your Microsoft account now (https://t.co/9oYF5zLe78). mine was there and it unlocked fine
• check Task Scheduler and your Startup folder to see what runs at login
what's the scariest thing your PC has done that was fine? 👀
tried out @GoEntire, the new hosting platform, and tbh I think I'm too locked in to switch lol
my whole setup lives on Cursor, Cursor Origin and cursor cloud agents, and I've been building on it for years. even when something else looks better, I can't give up the features
the big one for me: 𝗖𝘂𝗿𝘀𝗼𝗿 𝗰𝗹𝗼𝘂𝗱 𝗮𝗴𝗲𝗻𝘁𝘀
you could work around it, but you'd have to spin up a whole CLI every time you launch a cloud agent. way slower
are you locked into your stack too? 👀