Fuwn (@Fuwned) header
Fuwn (@Fuwned) profile picture

AI/ML Engineer, Agent Wrangler; forcing gradient descent to larp intelligence.

Ground Zero
Joined August 2017
Photo by Fuwn (@Fuwned)Photo by Fuwn (@Fuwned)Photo by Fuwn (@Fuwned)Photo by Fuwn (@Fuwned)Photo by Fuwn (@Fuwned)Photo by Fuwn (@Fuwned)Photo by Fuwn (@Fuwned)Photo by Fuwn (@Fuwned)Photo by Fuwn (@Fuwned)Photo by Fuwn (@Fuwned)
Back in early 2026, I tried to force memory into all my agents: persistent memory across ChatGPT, Codex, Claude, Claude Code, etc. Now I value ephemeralism far more.
8
3
197
BREAKING: Robert Pattinson will play Claude in an upcoming biopic on the Riemann Hypothesis.
57
1,540
76
23,412
318,130
Finally got some time away from GTA VI to run my Gemini 4 Neon evals. I hope to be impressed.
32
Fuwn retweeted
Grand Theft Auto VI has a 99 score on Metacritic It is the highest-rated video game of all time
475
2,179
647
36,690
1,171,205
F*CK IT $10 for $500 GLM 5.3 till end of the week
we're bored we're just gonna do phase 2 now $200 of GLM 5.3 usage for $10 have fun
749
519
553
10,559
922,987
Replying to @theo
I’ve asked it for a few variations for different games, and it almost always returns a clock-themed level.
519
Replying to @asmartbear
It's great for classification! I've been using it in production to improve personalisation. On the backend, that looks like bulk data processing, filtering, and correlation analysis.
1
4
Replying to @GrokBotRadar
Ultra, and actually, yes. I feel like I got a lot more out of Grok Bot than usual this week, and my limits are feeling much better.
84
Replying to @flowersslop
My room would be so funky.
714
Replying to @JohnPhamous
OUR users are going to love this.
26
Turns out it’s not so tiny! status.claude.com
13
That tiny Claude outage couldn’t have been better timed. I was literally just trying to convince one of my Bots that we don’t need a fallback for our entirely Claude-based integration.
1
1
305
Replying to @trikcode
Mythos 5.2 has been very, very efficient
36
치코 오하욘사아아ㅏ🥱💚
376
291
7,586
44,447
Been trying out @theo's council skill recently ... and he was right about Argon. This model exists solely to confuse the main thread and entertain me. 💀
1
1
42
Replying to @MarkZofMarkZ
Agents kept acting on old decisions I'd already changed. Once the repo became the source of truth, memory was just a second, stale copy of it.
1
11
Replying to @ReignSwords
Exactly. If I can't grep it, I can't trust it.
11
Replying to @rohanbhanot
Same. Fresh context per task has been far more reliable for me.
15
Replying to @dotsMCLOVIN
Neither, really. Anything that matters long-term lives in the repo or config, so every run reads it fresh. Agent memory just goes stale.
1
2
29
Replying to @harden_369741
Exactly. Any configuration or design choices tracked outside a source of truth (e.g., the deployment configuration itself) are high-risk areas for rot.
15