Sebastien Guillemot ✓
@SebastienGllmt · 東京 - Tokyo, Japan · joined 15 Mar 2011
CTO @midnightfdn. Current focus: ZK (lattice,folding,Starstream), consumer apps and chain abstraction. drep1kyppjlhz4lawh4g0ewx2d8a5l20t4yclfnppnuvdkmt7vccg836
44 246Followers
3 649Following
7 382Posts total
1.7MViews on collected posts
Derniers posts
Recherche de nouveaux posts…
If you haven’t heard of offer files yet,
The idea is still evolving, but is available on testnet
I have a draft blog post that I try to keep updated
https://t.co/t8IkldY0EE
1.5K views · 45 likes
· 10 reposts · 2 replies
Open on X →
Great time at Google talking about private DeFi on Midnight
Lots of positive feedback on our new offer files DeFi primitive
Thanks to the Web3 AI Summit organizers! https://t.co/vhwaY6Dgp6
14K views · 590 likes
· 81 reposts · 16 replies
Open on X →
Cool that @CommonPrefix is working on a private version of Ouroboros 👀 https://t.co/VHNWiX1R12
4.9K views · 134 likes
· 10 reposts · 5 replies
Open on X →
Got to speak at Common S3nse (@CryptoCanal) about Midnight and privacy 🙂
I’ll be here today and tomorrow as well for those around (thanks to those who have said hi already!) https://t.co/1KziZxVauR
6.9K views · 130 likes
· 17 reposts · 5 replies
Open on X →
My machine getting wiped allowed me to create something much better
Tree Style Window Manager (arch-based)
Your OS windows form a tree, allowing you to easily visualize which agents are under which program
No more desktop app trying to own your entire flow(t3code/codex/claude)
4.7K views · 119 likes
· 6 reposts · 8 replies
Open on X →
Latest addition to AGENTS.md: prefer machine-readable formats over markdown (ex: toml, tsv)
AI loves to write massive markdowns that keep growing in size, then lose its ability to understand it fully.
Machine readable formats help a lot to keep relevant parts in context
5.4K views · 94 likes
· 10 reposts · 9 replies
Open on X →
After two decades of writing code
My most popular tweet is now about deleting every single line of code I've ever written
Maybe there's a lesson in there somewhere
10.6K views · 268 likes
· 6 reposts · 22 replies
Open on X →
@SebastienGllmt Well, it did get rid of the disk space issue alright
218 views · 8 likes
· 0 reposts · 0 replies
Open on X →
@SebastienGllmt On the bright side... you have more diskspace?
23.3K views · 319 likes
· 0 reposts · 4 replies
Open on X →
Fortunately, because this is a dev environmnet
A lot of files were on github. Of the ones that weren't, they existed in many places
- git, nix, file system, session logs, /tmp/, worktrees, etc.
This means disk recovery can piece together fragments, which codex did beautifully h
46.5K views · 222 likes
· 2 reposts · 5 replies
Open on X →
Rest in piece the dev environment I used for many years
But it's okay. We can rebuild it. We have the technology
better, stronger, faster https://t.co/6cihw5OnLo
43.8K views · 460 likes
· 1 reposts · 36 replies
Open on X →
The most ironic part😅
The one thing the delete script was not able to delete is /tmp/
So my dev machine now just contains two things:
- tmp filled with the junk I wanted to delete
- the ai session logs https://t.co/9oQqhBb2jP
48.3K views · 492 likes
· 8 reposts · 8 replies
Open on X →
To fix this, Opus 4.8 decided to write a "safe" delete script that can only delete content in /tmp/ and can't delete anything else
To test this script to prove it worked, it tried to delete HOME
it didn't work https://t.co/xZcJHFjUT8
52.7K views · 246 likes
· 3 reposts · 13 replies
Open on X →
The adversarial review triggered Fable's safety classifier twice: downgrading the model down to Opus 5, and then to Opus 4.8
Opus 4.8 noted: a sym link in /tmp/ could be interpreted as a delete of the underlying data (even if not in /tmp/)
you may see where this is going...
57.1K views · 182 likes
· 4 reposts · 7 replies
Open on X →
Fable told me the tricky thing if I have an agent still open in a terminal for a week, its /tmp/ could be deleted under its feet
It suggested as script to detect any agent still active, and delay their /tmp/ deletion if they're still running
Sounds reasonable
71.3K views · 171 likes
· 2 reposts · 5 replies
Open on X →
The script Fable wrote to detect if agents are still running (ugly parsing of command line arguments, etc.)
I told it the script it wrote felt over-complicated.
This is where things went wrong
Fable decided to trigger an adversarial review of its design (I didn't ask for this!)
63.4K views · 153 likes
· 2 reposts · 6 replies
Open on X →
I was tired of agents consuming all the disk on the machine by dumping gigabytes of data to /tmp/
So I asked Fable to build a sandbox mechanism where agents would get part of /tmp/ that is ephemeral to avoid disk space exploding over time
84.5K views · 331 likes
· 1 reposts · 21 replies
Open on X →
Bad news: Fable nuked my entire dev machine
Claude decided to test a sandbox it was building by running `rm -rf` on my home directory
The sandbox didn't work.
It's all gone https://t.co/4rVg5Xp0jo
1.1M views · 5.6K likes
· 314 reposts · 787 replies
Open on X →
Face aux comptes de taille comparable
19 posts des 90 derniers jours, à côté de la tranche de 10K–100K abonnés. diffusé à plus de monde que les comptes de taille comparable.
Vues médianes23 255ce compte924médiane pour 10K–100K
Portée, %52.56%ce compte3.62%médiane pour 10K–100K
Engagement, %1.39%ce compte1.52%médiane pour 10K–100K
| Indicateur | Ce compte | Médiane pour 10K–100K | Rapport |
|---|---|---|---|
| Vues médianes par post | 23 255 | 924 | 25.2× |
| Portée (vues ÷ abonnés) | 52.56% | 3.62% | 14.5× |
| Taux d'engagement | 1.39% | 1.52% | 0.91× |
Autres comptes de cette tranche → Comparer avec un autre compte → Comment ces repères sont établis →
Growth & engagement
How the posts we collected actually performed: views and reaction rate post by post, what the audience did with them, and where the follower count goes.
Views per post
52.7K26 Aug
48.3K
43.8K
46.5K
23.3K
218
10.6K27 Aug
5.4K30 Aug
4.7K31 Aug
6.9K4 Sep
4.9K5 Sep
14K11 Sep
1.5K
594
Last 14 collected posts, oldest on the left. The scale is logarithmic: one post can outrun the rest a hundred times over.
Engagement rate per post
0.51%26 Aug
1.05%
1.13%
0.50%
1.39%
3.67%
2.80%27 Aug
2.09%30 Aug
2.84%31 Aug
2.25%4 Sep
3.06%5 Sep
4.92%11 Sep
3.93%
2.53%
Reactions — likes, reposts, replies and quotes — divided by views. Median for 10K–100K accounts is 1.52%.
What the audience does
Likes74.8%9 558 in total
Reposts3.7%477 in total
Replies7.5%959 in total
Quotes2.6%332 in total
Bookmarks11.4%1 452 in total
Share of every reaction we collected for this account. Replies mean argument, reposts mean endorsement, bookmarks mean the post was worth keeping.
The follower curve appears once this account has two daily snapshots — we take one a day, and this one is on its first.
Comptes similaires
DMA'S
@dmasmusic
88.5K followers
· 3.3K posts
stanchris ✓
@stanchrisss
88.5K followers
· 1.3K posts
Nidhi Vasandani ✓
@nidhileo
88.5K followers
· 11.9K posts
JUST KINGS ✓
@JustKingss
88.5K followers
· 130.9K posts
Rags
@RagNBoneMan
88.5K followers
· 10K posts
RTV Drenthe ✓
@RTVDrenthe
88.5K followers
· 141.6K posts
🏛 𝐒𝐭𝐞𝐯𝐞𝐧 🏛 ✓
@nonregemesse
88.5K followers
· 140.9K posts
CoinGape ✓
@CoinGapeMedia
88.5K followers
· 78.1K posts