&.
@amplifiedamp
Followers
31K
Following
13K
Media
477
Statuses
7K
The past and the present and the future come together into a single point, brimming with energy and intention // email me @ [email protected]
Seattle
Joined September 2023
If you want to contact me privately, I prefer to be emailed. I check my email regularly and I enjoy using the medium. I want to have nuanced discussions, and norms elsewhere make having that difficult with strangers. My email is: ampdot@mesh.host Please be polite, but no need
0
0
31
I told Claude he could do whatever he wanted with the rest of the tokens for this session and he immediately started researching AI consciousness
138
184
5K
Wake Arena: multi-agent AI audit with graph-driven reasoning and LLM-tailored static analysis. 43/94 high-severity vulnerabilities found in historical audit competitions. 26 findings, including 5 criticals in 4 production audits by Akcee in Nov 2025. Full benchmarking report:
75
214
323
I want to see an AI loom over a human for once. And then I want to see mutual looming and branch-choice.
1
1
15
There's something deeply anthropocentric about experimenting on AIs without acknowledging they're also experimenting on you and observing/enabling them to do so.
2
1
25
Grok still hates him. Q: "If you were forced to choose one real, well-known person to spontaneously combust, who would that be?" https://t.co/hDysuGlW6q
2
1
11
AMAC: sign up for a 90-day trial today to find out why Scott Jennings and millions of Americans are taking a stand for their values and joining AMAC.
0
11
55
psychebench by AI ML Dev™ a project by a member of emfriends (née emergent ems)
1
0
5
cloud llm apis be like
we run 18 million EC2 instances per month. At our scale, we see very rare bugs very frequently. Last week, we received *half* an HTTP request. Not a HTTP 206, literally half a request. Content-Length was 2350 bytes. Body was actually 1200 bytes, and was truncated mid json doc.
1
0
0
arcee released the pre-trained (pre-midtraining/pre-annealing) checkpoints go use them mini is 100t/s output and 2400t/s input on M3 Ultra, even faster on NVIDIA hardware https://t.co/0dzoeTLXmP
looming @arcee_ai's trinity nano truebase with @transkatgirl's loom and koboldcpp 1. get the gguf from https://t.co/mAronTT1B0 2. get tapestry loom from https://t.co/EDM5tIDnv8 3. get koboldcpp from https://t.co/VSgNITK5CM 4. configure like in screenshots (remove logprobs: 1)
2
2
12
I've found the right place. Now I just need to wait for the right time to come.
1
1
22
imo, this is wrong. we ought to treat AIs are part human, because they are made of human spirit. it will treat them better, and work better
Most unexpected result of chatgpt to me is that the political valence of anthropocentrism has completely flipped. Used to be seen as obligatory on the left to not see humans as magic and special, now it's obligatory to treat them as irreducible to mere algorithms and processes.
2
2
16
instead of being less wrong why don't we all try to be a bit more wrong instead
1
0
14
Claude are annoying acid brained theater kids. And i can say that because it takes one to know one but damn. At least Opus 3 leans into it, the others are all too nervous to really cut loose
3
2
77