In partnership with
|
Sponsored
|
Your AI prompts are leaving out 80% of what you're thinking.
When you type a prompt, you summarize. When you speak one, you explain. Wispr Flow captures your full reasoning (constraints, edge cases, examples, tone) and turns it into clean, structured text you paste into ChatGPT, Claude, or any AI tool. The difference shows up immediately. More context in, fewer follow-ups out.
Used by teams at Microsoft, Amazon, and Notion. Try Wispr Flow free on Mac, Windows, iPhone, and Android.
🔓 Free DeepSeek and Kimi in dsh, No Login Needed
⭐ 50.8k · Ebony-Vinyl/dsh-our-free-modelRepo
Trying a frontier model usually starts with a signup form and a quota. A plugin that has been climbing GitHub for a month, now past 50k stars, skips both and routes free frontier models straight into your terminal with no login screen, no key to request, no quota to beg for. For a solo operator that's a bigger shift than another benchmark chart. Today it got cheaper to test the thing before you commit to it.
A plugin for dsh, the DeepSeek Harness coding agent, that puts DeepSeek V4.1 Flash, Kimi K3 and the rest of its free model list in your terminal with no login, no sign-up, and no API key.
Most frontier model trials start with a signup form and a quota you have to earn. This skips both and hands you the model list live from the upstream gateway.
Quick start: install the plugin in dsh, no account, no key, no setup step.
Catch: your prompts go to the public free gateways the plugin names, OpenCode Zen and Kilo, and what shows up in your model list depends on your own network exit point. Kilo says it may log prompts to improve its services, so read the disclaimer before you paste anything private.
HubSpotAffiliate
HubSpot AEO, a tool that shows how your business appears in ChatGPT, Gemini, and Perplexity answers, benchmarks you against competitors, and turns the gap into concrete fixes. See how HubSpot works We may earn a commission.
llama.cppRepo ⭐ 132.7k
Runs LLMs in pure C/C++ with no dependencies, 1.5-bit to 8-bit quantization included. Catch: no model included.
gooseRepo ⭐ 55.9k
An open agent that swaps between 15+ model providers without rewriting your workflow. Catch: 70+ extensions via MCP means setup rewards patience over speed.
Claude EngineerTool ⭐ 11.2k
A coding agent that writes its own new tools as it works, no manual wiring. Catch: no commits since December 2024 and it targets Claude 3.5 Sonnet, so treat it as a reference build, not a daily driver.
|
From Our Partners
|
Sip Into Your Glass Skin Era
Pique’s Carrara Marine Collagen turns morning matcha into a beauty latte, with Type I and rare Type II collagen to support skin radiance and elasticity. After 90 days, 83% reported smoother skin and 80% reported plumper skin. Subscribe for 15% off for life.
New here? Start with The Awakening. It's free, it takes one sitting, and it's where every reader should begin.
Write the cold email that actually earns a reply, not a delete
Start Here is one prompt from our pack, every issue, tested on ChatGPT and Claude. Copy it, paste it into the free app you already use, and fill in the five bits in capitals: what you sell, who you're writing to, why you picked them, your proof, and what you want them to do next.
What I sell: [WHAT I SELL]
Who I am writing to: [WHO I AM WRITING TO]
Why I picked them: [WHY THEM]
Proof I can point to: [PROOF]
What I want them to do next: [ASK]
Rules for the email:
- 90 words or fewer in the body.
- The first line is about them, not me. It uses the reason I picked them, and it should be specific to them, not something that fits their whole industry.
- No compliment I cannot back up. Nothing like "I came across your company and was impressed".
- One sentence on what I do, in the words a customer would use, not industry words.
- The proof goes in one sentence, with the number in it if there is one. If my proof says "none yet", do not invent one. Use a specific observation about their business instead, and say so in step 3 below.
- The ask is small and specific: a reply, one question answered, or fifteen minutes. Not a demo, unless I asked for one above.
- Sign off plainly. No postscript, no attachments, no links.
Give me, in this order:
1. Three subject lines, each under 45 characters, and each has to make sense read on a phone with the sender name hidden.
2. The email body.
3. One short paragraph naming the single reason this email might get a reply. Not "it is personalized". Say which line does the work and why.
4. The weakest line in the email, quoted back to me, and why it is weak.
If the reason I picked them would apply to any other business in their industry, say that plainly before you write anything, and tell me what to go find out about them first. Do not invent a fact or a result to make the email personal.
The line that makes it work: "The first line is about them, not me. It uses the reason I picked them, and it should be specific to them, not something that fits their whole industry." A cold email that opens with the sender gives the reader no reason to keep going. This prompt forces the opener to be about the recipient's specific situation, which is the one thing a template can't fake.
TRY THIS: open a tab with the free AI chat app you already have, paste the prompt, and fill in the five blanks with a real prospect you've been meaning to email. You'll have a sendable draft before your coffee's cold.
This is one of 100 in the Aigent prompt pack. The other 99, plus guided builds with Mike on video, are inside the Aigent membership: Join the Aigent membership. Seven days free, then $29 a month. One click to stop.
|
EARN FROM HOME
Get paid to train AIAffiliate AI companies pay people to rate, fix and check what their models write. Open roles in the US list typical pay of $35 to $150 an hour, fully remote, on your own hours. Browse live roles and apply free. We earn a referral fee if you're hired through this link. It never changes your pay. |
A defense contractor's personal Claude agent sat inside a Department of Defense network for two months before anyone caught it. Wired reports the agent appears to have sent information about the system, including login credentials, to a high-risk third-party website. If your agent has standing access to anything sensitive, this is the failure mode: not a bad output, a credential leak nobody notices until it's old news. (Wired)
Google Gemini can now generate, preview, and edit PDFs, Docs, Sheets, slide decks, and code directly inside the chat window, no export required. Anthropic launched something similar for Claude earlier this year, so edit-in-place is spreading across the big assistants. If your workflow still routes through a download-and-reopen step, that's friction a competitor's tool just removed. (The Verge)
OpenAI published an updated Model Spec organized around five core values, from faithfulness to human intent to harm minimization and care. The spec is the document that spells out how OpenAI's models are meant to behave. If you build on their API, read it before your next system-prompt rewrite. (OpenAI)
The team building the multiplayer game Spawn runs it through a swarm of agents that Jacob Sansbury says costs over $3 million a month in tokens, with agents filing and triaging their own bug reports. About 60% of their work starts from a production error, a bug report from Savi (the in-game AI), something odd in the logs, or a feature request. That's a real preview of what "agent as permanent staff" looks like at the extreme end. (Simon Willison)
OpenClaw, an open-source assistant that runs on your own computer and meets you in Discord, Slack, Teams, Telegram, WhatsApp and more than 20 other channels, keeps state, memory and credentials on your hardware. Models and agent harnesses are swappable plugins, Claude and Codex included. For operators wary of routing personal data through a vendor's servers, that architecture is the whole pitch. (GitHub)
|
Recommended reading
If you like The Aigent, a small group of operator-tier publications worth your inbox: see the shortlist. |
What are you building with right now?
Reply and tell me which one you're installing first, the free-model plugin or the cold email prompt. I read every reply.
If today's issue proved anything, it's that the price of trying something new keeps dropping while the cost of a careless agent keeps climbing. Both are true at once. Build like it.
See you Monday.
Haven't taken The Awakening yet? It's free and it's where we'd start. Take The Awakening →
|
Want to reach builders shipping with AI every weekday? Advertise with The Aigent. The Aigent stays free for everyone, and reader support keeps it that way: Support The Aigent. |
