Re: [NonGNU] new: hermes

>> While Hermes Agent can be run locally, its development, documentation,
>> and promotional materials are overwhelmingly geared toward a "service 
>> as
>> a software substitute" (SaaSS) model, which centralizes control with 
>> the
>> service provider and denies users agency over their own computing.
> 
> But this paragraph takes the actual usage pattern, and makes it the
> design tendency.  What is the evidence for that?  Do you see any
> features that discourage usage of free LLMs locally? if so, what are
> those discouraging features and how do they affect users who want and
> do use local LLMs?

Yes, that is design tendency.

Hermes as software agent cannot be installed without registering with 
the "portal", see below killed text from console install process:

→ Starting setup wizard...


┌─────────────────────────────────────────────────────────┐
│             ⚕ Hermes Agent Setup Wizard                │
├─────────────────────────────────────────────────────────┤
│  Let's configure your Hermes Agent installation.       │
│  Press Ctrl+C at any time to exit.                     │
└─────────────────────────────────────────────────────────┘

     Skipped (keeping current)

◆ Nous Portal
   One subscription, 300+ models, plus the Tool Gateway:
     web search, image generation, TTS, browser automation.
   Sign up: https://portal.nousresearch.com/manage-subscription

Not logged into Nous Portal. Starting login...

Starting Hermes login via Nous Portal...
Portal: https://portal.nousresearch.com

To continue:
   1. Open: 
https://portal.nousresearch.com/manage-subscription?user_code=AMH4-5SDB
   2. If prompted, enter code: AMH4-5SDB
   (Opened browser for verification)
Waiting for approval (polling every 1s)...

Then once you access that website, user is of course directed onto the 
"plans" page to subscribe possibly to service as a software substitute 
LLMs.

>> By promoting Hermes Agent, the GNU project would be endorsing a system
>> that normalizes dependency on non-free network services, undermining 
>> the
>> very principles of freedom and self-determination it was founded to
>> protect.
> 
> And this is another "leap of faith".  We have no control on what our
> users do, we just should not actively encourage them to do the wrong
> thing.  If a package supports both free and non-free backends, I see
> no reason to claim that its support for non-free ones is the
> overwhelming trait, let alone that it causes us endorsing of such
> backends.

Nobody should control users. That was whole point from my side.

Enabling Hermes is actively encouraging users to subscribe to Hermes 
deals using Service as a Software Substitute LLMs.

Surely users can get something "free" like run local models.

That one thing is free and other 200 things are not-free seems to be now 
justified.

Yes, the support for non-free ones is the overwhelming trait. So far on 
this list there is nobody who even mentioned of running local models. 
But we have people mentioned running Claude.

The full setups says:

Select provider:
   ↑↓ navigate  ENTER/SPACE select  ESC cancel

  → (●) Nous Portal (Everything your agent needs, 300+ models with 
bundled tool use)
    (○) Fireworks AI (OpenAI-compatible direct model API)
    (○) OpenRouter (Pay-per-use API aggregator)
    (○) Mixture of Agents (named presets; aggregator acts after reference 
models)
    (○) NovitaAI (Cloud: Model API, Agent Sandbox, GPU Cloud)
    (○) LM Studio (Local desktop app with built-in model server)
    (○) Anthropic (Claude models via API key or Claude Code)
    (○) OpenAI ▸ (Codex CLI or direct OpenAI API)
    (○) Qwen ▸ (Qwen Cloud / DashScope, Coding Plan & Qwen CLI OAuth)
    (○) xAI Grok ▸ (Direct API or SuperGrok / Premium+ OAuth)
    (○) Xiaomi MiMo (MiMo-V2.5 and V2 models: pro, omni, flash)
    (○) Tencent TokenHub (Hy3 Preview via tokenhub.tencentmaas.com)
    (○) NVIDIA NIM (Nemotron models via build.nvidia.com or local NIM)
    (○) GitHub Copilot ▸ (GitHub token API or copilot --acp process)
    (○) Hugging Face Inference Providers
    (○) Google AI Studio (Native Gemini API)
    (○) Google Vertex AI (Gemini via GCP; OAuth2 service account or ADC, 
GCP billing/quotas)
    (○) DeepSeek (V3, R1, coder, direct API)
    (○) Z.AI / GLM (Zhipu direct API)
    (○) Kimi / Moonshot ▸ (Coding Plan, Moonshot global & China 
endpoints)
    (○) StepFun Step Plan (Agent / coding models via Step Plan API)
    (○) MiniMax ▸ (Global, OAuth Coding Plan & China endpoints)
    (○) Ollama Cloud (Cloud-hosted open models, ollama.com)
    (○) Arcee AI (Trinity models, direct API)
    (○) GMI Cloud (Multi-model direct API)
    (○) Kilo Code (Kilo Gateway API)
    (○) OpenCode ▸ (Zen pay-as-you-go or Go subscription)
    (○) AWS Bedrock (Claude, Nova, Llama, DeepSeek; IAM or API key)
    (○) Azure Foundry (OpenAI-style or Anthropic-style endpoint, your 
Azure AI deployment)
    (○) Vercel AI Gateway (Multi-model aggregator)
    (○) Actual Computer - hosted inference via api.actual.inc, or local 
offline inference via ACTUAL_BASE_URL
    (○) custom (direct API)
    (○) DeepInfra — 100+ open models, pay-per-use
    (○) Upstage (Solar API)
    (○) Custom endpoint (enter URL manually)
    (○) Configure auxiliary models...
    (○) Leave unchanged

Commercial/Cloud/SaaS providers	is about 30 of them (Nous Portal, 
Fireworks, OpenRouter, Anthropic, OpenAI, Google, AWS, Azure, etc.), 
while 2 options for locally running LLM

That's a ~30:2 ratio of cloud/commercial to local options. "Nous Portal" 
is pre-selected (the ●)

It is overwhelming to me. The "Actual Computer" option combines hosted 
and local inference but defaults to hosted. The setup flow forces portal 
registration before the software can be used, with Nous Portal 
pre-selected as the default. This is not a neutral tool — it is designed 
to funnel users into SaaSS subscriptions, with local/free use as a 
neglected edge case. Including Hermes in GNU would actively promote this 
funnel.

It violates the GNU spirit of free software by making free use 
impractical, steering people to non-free SaaS LLMs.

-- 
Jean Louis