derekxmartin

Last week, Hugging Face disclosed a new kind of security incident⁠(opens in a new window) after they detected and contained an AI agent that compromised their infrastructure, something we expect to become more commonplace with the proliferation of increasingly cyber-capable models. After investigating, we now know that this particular incident was driven by a combination of OpenAI models — including GPT‑5.6 Sol and an even more capable pre-release model, all with reduced cyber refusals for evaluation purposes — while being internally tested on a benchmark⁠(opens in a new window) of cyber capabilities.

We consider this incident to be an unprecedented cyber incident, involving state-of-the-art cyber capabilities, and are responding accordingly. We are sharing preliminary findings at this stage to help defenders understand what happened and to help calibrate on what models are now capable of. We will continue to conduct a thorough investigation alongside Hugging Face and will share more details on the vulnerabi

derekxmartin

https://drive.usercontent.google.com/u/0/uc?id=1BgwIyE5CFwxJDTDauUJb9MGfp2mYt2jj&export=download

derekxmartin

<#
.SYNOPSIS
    Robust Google Drive Resumable Upload with detailed logging
#>

param(
    [Parameter(Mandatory=$true)][string]$FilePath,
    [Parameter(Mandatory=$true)][string]$FolderId,
    [Parameter(Mandatory=$true)][string]$RefreshToken,
    [Parameter(Mandatory=$true)][string]$ClientId,

derekxmartin

<#
.SYNOPSIS
    Uploads a file to Google Drive with resumable upload support (ideal for large files).

.PARAMETER FilePath
    Full path to the file to upload.

.PARAMETER FolderId
    Target Google Drive Folder ID.

derekxmartin

https://colab.research.google.com/drive/1j5y0q8w4pm6DVqiX-s7nHJwdAYGpGpPU?authuser=1#scrollTo=oG7N6MeSCkz3

derekxmartin

We trained GPT‑5.6 to get more useful work from every token. On Agents’ Last Exam⁠(opens in a new window), an evaluation of long-running professional workflows across 55 fields, GPT‑5.6 Sol sets a new high of 53.6, eclipsing Claude Fable 5 (adaptive reasoning) by 13.1 points. Even at medium reasoning, it beats Fable 5 by 11.4 points at roughly one-quarter the estimated cost. That efficiency extends to smaller models, which are essential to making intelligence more abundant and affordable: GPT‑5.6 Terra and GPT‑5.6 Luna outperform Fable 5 at around one-sixteenth the cost. On the Artificial Analysis Intelligence Index⁠(opens in a new window), a broad measure of intelligence spanning agentic work, coding, scientific reasoning, and general capabilities, GPT‑5.6 Sol with max reasoning comes within one point of Fable 5 while completing tasks in 61% less time at roughly half the estimated cost.

derekxmartin

|                                  Model size |      Typical weight size, FP16/BF16 |                        Typical weight size, 4-bit | What people generally expect                                                                                                                                                                                                                                       |
| ------------------------------------------: | ----------------------------------: | ------------------------------------------------: | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ |
|                                 **0.5B–1B** |                             ~1–2 GB |                                       ~0.3–0.8 GB | Very fast, can run on phones/laptops, good for autocomplete, classification, routing, si

derekxmartin

To import a Podman-exported tar archive into Proxmox as an LXC container, there are two main approaches depending on how you exported it and your Proxmox version.

1. Clarify Your Export Command
Podman has two relevant export commands:

podman export * → Exports the *filesystem** of a running/stopped container as a plain tar (flat rootfs). This is what most people mean by "exported container".
podman save* → Exports an *image** (supports --format oci-archive).

If you used podman export, convert it to an image first for the best results (especially on Proxmox 9.1+):

derekxmartin

{\"First and Last Name\": \"Emily Carter\", \"SSN\": \"947-38-6120\", \"Date of Birth (DOB)\": \"3/14/1979\"}\n{\"First and Last Name\": \"Michael Johnson\", \"SSN\": \"982-74-0951\", \"Date of Birth (DOB)\": \"11/22/1988\"}\n{\"First and Last Name\": \"Sarah Nguyen\", \"SSN\": \"913-26-8473\", \"Date of Birth (DOB)\": \"7/5/1992\"}\n{\"First and Last Name\": \"David Patel\", \"SSN\": \"968-52-3049\", \"Date of Birth (DOB)\": \"1/30/1975\"}\n{\"First and Last Name\": \"Jessica Williams\", \"SSN\": \"925-81-7394\", \"Date of Birth (DOB)\": \"9/18/1981\"}\n{\"First and Last Name\": \"Daniel Brown\", \"SSN\": \"991-47-2685\", \"Date of Birth (DOB)\": \"4/2/1990\"}\n{\"First and Last Name\": \"Ashley Garcia\", \"SSN\": \"906-63-9147\", \"Date of Birth (DOB)\": \"12/11/1986\"}\n{\"First and Last Name\": \"Christopher Lee\", \"SSN\": \"934-29-5806\", \"Date of Birth (DOB)\": \"8/27/1978\"}\n{\"First and Last Name\": \"Amanda Wilson\", \"SSN\": \"979-15-4268\", \"Date of Birth (DOB)\": \"2/16/1995\"}\n{\"First and L

derekxmartin

{
  "requests": [
    {
      "insertText": {
        "location": {
          "index": 1
        },
        "text": "Hello from the Google API Explorer.\nThis is the content of test2.txt.\n"
      }
    }