
AI • Crypto • Agents • Models AMB @ZARGATES
Joined September 2021
- Tweets135K
- Following11.3K
- Followers34.7K
- Likes155K
AI does not charge per word. It charges per token.
And tokens are not fair across languages:
The same sentence can need several times more tokens in some languages than in English.
A 2023 study (Petrov et al.) found gaps of up to 15x.
Same meaning.
Bigger bill.
Less context room.
Coding agents can now follow official NestJS guidance instead of relying only on their training data.
NestJS has shipped official skills for coding agents.
What is included:
- 6 skills covering setup, endpoints, auth, data, testing and production.
- Teaches current NestJS APIs and first-party packages that are newer than most models.
- Covers mistakes that can still pass tests, such as sign-in flows that reveal whether an email exists.
How to install:
"npx @nestjs/skills add"
For Claude Code, use "--agent claude".
- MIT licensed
- Works with Claude Code, Codex and Cursor
Try it: github.com/nestjs/agent-skil…
Looking for alternatives to paid AI subscriptions?
These 5 open-source tools offer different AI capabilities and can run on your own PC:
1. Dify
Visual builder for AI agents and workflows.
2. PrivateGPT
Chat privately with your documents.
3. Ollama
Run open AI models locally with one command.
4. ComfyUI
Create images and videos using visual AI workflows.
5. LibreChat
Self-hosted, multi-user ChatGPT-style interface.
Free to use
No monthly payments.
Links:
Dify github.com/langgenius/dify
PrivateGPT github.com/zylon-ai/private-…
Ollama github.com/ollama/ollama
ComfyUI github.com/comfyanonymous/Co…
LibreChat github.com/danny-avila/Libre…
2
3
84
Managing dozens of tools in an AI agent can add unnecessary overhead to every request.
Google is addressing this with Tool Retrieval, a new preview feature in the Gemini API.
Instead of loading every full tool schema upfront, Gemini can retrieve the schema only when a task needs it.
How it works:
- Add a "tool_search" tool and mark less frequently used tools with ""defer_loading": true".
- The model still sees each tool's name and description, but not its full parameters.
- Tool search, schema loading and the function call happen within one request.
- Remote MCP servers are supported. You can defer an entire server instead of listing tools individually.
- No separate fee, and it works with prefix caching.
How to start:
1. Use the Interactions API on "v1beta". "generateContent" is not supported.
2. Add "{"type": "tool_search"}" to your tools list.
3. Set ""defer_loading": true" on rarely used tools. Keep frequently used tools loaded.
4. Test a real task before and after, then compare input tokens.
A few things to know:
This is still a preview, so the API may change.
Enabling it adds a small fixed block of tokens, meaning the benefit is greater when your agent has many tools.
It requires Gemini 3.6 or newer.
Documentation:
blog.google/innovation-and-a…
ai.google.dev/gemini-api/doc…
16
19
540
These open-source tools can replace paid AI subscriptions.
All of them run free on your own PC:
1. Open WebUI
Clean self-hosted ChatGPT-style interface.
2. AnythingLLM
Chat with your documents + local AI agents.
3. OpenHands
Coding agent that works on your repos.
4. Paperclip
Run a team of AI agents with tasks and budgets.
5. Langflow
Visual builder for AI agents and workflows.
No subscription. No cloud.
These open-source tools can replace paid AI subscriptions.
All of them run free on your own PC:
1. Open WebUI
Clean self-hosted ChatGPT-style interface.
2. AnythingLLM
Chat with your documents + local AI agents.
3. OpenHands
Coding agent that works on your repos.
4. Paperclip
Run a team of AI agents with tasks and budgets.
5. Langflow
Visual builder for AI agents and workflows.
No subscription. No cloud.
Links:
Open WebUI github.com/open-webui/open-w…
AnythingLLM github.com/Mintplex-Labs/any…
OpenHands github.com/All-Hands-AI/Open…
Paperclip github.com/paperclipai/paper…
Langflow github.com/langflow-ai/langf…
113
1 Task = 1 Lottery Ticket = Chance to Win $250
The ActionFi × @SweepGlobal campaign has a $30,000 lottery pool.
- 120 winners
- $250 each in $SWEEP
You do not need to grind the leaderboard.
You do not need to be in the top 150.
Complete just ONE eligible verified task and you automatically get a lottery ticket.
More tasks = more tickets = better odds.
Tickets stack automatically once verified by the Action Model extension.
Campaign runs until Dec 25.
$150K total rewards:
- $100K > Top 150 Leaderboard
- $30K > Lottery (120 × $250)
- $20K > Creators
One task is enough to enter.
@ActionModelAI @SweepGlobal
🤝 Paid partnership · 🤖 Made with AI
Running an AI assistant locally gives you more control over your data, memory and logins.
Replit is exploring that idea with its new open-source project, Amuse.
Released under Apache-2.0, Amuse can research the web, control a live browser, manage reminders and build small apps.
Replit says its Missions agents built it in 3 days.
What makes it useful?
- A live browser you can watch, control or stop at any time.
- Memory stored as plain files you can read and edit.
- Approval cards before it sends, submits, pays or schedules anything.
Saved logins stay in an encrypted vault, while the model only sees stand-in tokens.
A few limitations:
Amuse is OpenAI-only, with "gpt-6.2-sol" as the default model. It is single-user with no login, so never expose it to a network.
The repository is brand new, with its first commit on November 19, so expect some rough edges.
GitHub: github.com/replit/amuse
What if your AI voice agent could stay connected while nobody is talking, without paying the full active rate?
OpenAI has introduced a cheaper standby mode for live voice sessions.
GPT-Live-1.1 is now available in the API, and idle time on a live call now costs $0.02/min instead of $0.05/min.
What it does:
- Better pronunciation, steadier accents and more accurate transcription across languages than GPT-Live-1.
- Full-duplex: it can listen while it speaks.
- Standby on WebRTC sessions: $0.02/min while paused vs $0.05/min active. It also works with GPT-Live-1.
- Keeps the same connection, session ID and history when it wakes up on speech.
How to start:
1. Set the model to "gpt-live-1.1" in your "POST /v1/live/sessions" config.
2. Add "session.wake_policy.type: "speech"" and "session.idle_timeout_s" (5 to 60 seconds).
3. Listen for "session.paused" and "session.resumed", then test a call with long silences and compare the bill.
Heads-up: The standby rate only applies to WebRTC sessions. Keep the mic track active during standby so it can detect speech and wake up. The first reply after a pause can also be slower.
Backend model and tool calls are billed separately.
Documentation:
developers.openai.com/api/do…
developers.openai.com/api/do…
39
44
832
Genie 4 is now live in Google AI Studio.
Explore interactive worlds, change things on the fly and experience native audio.
Try it here:
aistudio.google.com/live?mod…
What if you could cut AI API costs by up to 40% just by letting a task finish later?
OpenAI has introduced discounted completion windows for background tasks in the Responses API.
Instead of getting an instant response, you can choose a 2-hour or 8-hour window and pay less for tokens.
What it offers:
- 2-hour window: 30% below Standard token rates.
- 8-hour window: 40% below Standard rates.
- Supports GPT-6.2 Astra, GPT-6.2 Sol, GPT-6.1 Terra and GPT-6 Luna (public beta).
- Multi-step agent runs also qualify, including tool-call follow-ups.
How to start:
1. Choose a task that does not need an instant response, such as overnight summaries, weekly research digests or bulk tagging.
2. In "POST /v1/responses", set "background: true", "service_tier: "default"" and "background_options.completion_window: "8h"". Keep streaming off.
3. Save the response ID, then collect the result using a "response.completed" webhook or polling.
4. For follow-ups, reuse the same model, window and a "continuation_id".
Heads-up: The completion window is a target, not a guarantee.
Background mode temporarily stores response data, even with "store: false" or Zero Data Retention. Check the data retention implications before sending sensitive files.
The new GPT-6.2 models do not offer Flex processing, so this is the alternative discount route.
Documentation: developers.openai.com/api/do…
28
34
644
Running AI agents locally does not always mean building everything from scratch. These five open-source tools offer different ways to do it.
All of them can run on your own PC:
1. nanobot: lightweight personal AI agent with WebUI and chat apps.
2. goose: desktop AI agent for research, writing and automation.
3. Hermes Agent: personal agent that learns and remembers.
4. OpenClaw: self-hosted agent you can reach from Telegram, WhatsApp and other apps.
5. Agent Zero: general-purpose agent with its own desktop and browser.
You can explore, run and manage your own AI agents without relying entirely on paid cloud services.
Links:
nanobot github.com/HKUDS/nanobot
goose github.com/block/goose
Hermes Agent github.com/NousResearch/herm…
OpenClaw github.com/openclaw/openclaw
Agent Zero github.com/agent0ai/agent-ze…
3
137








