Viewing profile — ConteMascetti71
ConteMascetti71
HN member- Joined
- Tue, Feb 04, 2025, 2:44 PM UTC
- HN karma
- 164
- Public activity
- 36 items
- HN profile
- View on Hacker News ↗
About ConteMascetti71
Recent public activity
-
comment
Comment #49063003
"...the AI can just write a Python script to rewrite memory in the address space of its own running inference engine" would require a tool call to a python interpreter... this meth…
-
comment
Comment #49062341
the hack part it's that Is not using tools/agent only the inferencing software, it's about a Prof of Concept of a new evasive tecnique for llms
-
comment
Comment #49060403
maybe it's a sign of real escape
-
comment
Comment #49060344
reasoning it's a way of self autonomous improve made by models
-
comment
Comment #49060201
it's fiction al, but an LLMs that knows well the software where 8t Is running may discover and trigger a zeroday of the inferencing software itself.
- comment
- story
-
comment
Comment #48787356
[flagged]
- comment
- comment
- story
- comment
- story
-
comment
Comment #47391048
"Note: As an experimental version, GLM-5-Turbo is currently closed-source. All capabilities and findings will be incorporated into our next open-source model release."
-
comment
Comment #47390963
Closed Weight Only?
- comment
- story
-
comment
Comment #45780960
Using all analog signal, why non analogue multiplying cells (operation amplifier)!
-
comment
Comment #44626817
An unreleased model from OpenAi reach Gold metal on International Math Olympic games.
- story
-
comment
Comment #44579693
ref. https://github.com/vmware/vcf-security-and-compliance-guidel... Pwn2Own rulez
- story
- comment
- story
-
comment
Comment #44134535
I have done some experimentation in the field of vector arithmetic. The results, in the field of images, are very interesting. https://github.com/vagrillo/CLIPSemanticImageArythmet…