- Outlier Coder's Newsletter
- Posts
- Coder's Newsletter # 13
Coder's Newsletter # 13
š Gemini CLI lets you code, automate, and research via the terminal
Hey Outliers,
Welcome back to Outlier coderās Newsletter! š Your front-row seat to AIās biggest moves, smarter tools, and key research updates.
Some AI updates help you write code faster. Some help you work smarter. And someāunexpectedlyājust listen. This weekās news shows how AI is creeping into places we didnāt plan for, but now might not want to build without.
This Week in AI
Gemini CLI lets you code, automate, and research via the terminal
Anthropic explores Claudeās role in emotional support use cases
OpenAI enhances API with web search, webhooks, and deep research tools
š„ Whatās hot this week?š„
š Googleās Gemini CLI: AI-Powered Command Line for Developers

Google has launched Gemini CLI, an open-source AI agent that brings the power of Gemini directly into the terminal. Designed for developers, this tool streamlines coding, debugging, automation, and cloud management through natural language commands right from the command line.
Key Highlights:
Direct Terminal Integration: Use Geminiās AI models right in your terminal to code, debug, and automate with natural language commands.
Generous Free Tier: Enjoy free access to Gemini 2.5 Pro, with a large 1 million token context window and high daily and per-minute usage limits.
Versatile Functionality: Generate content, solve problems, research topics, fetch real-time documentation, deploy apps, manage cloud resources, and automate scriptsāall within your terminal.
Seamless Integration: Works with Gemini Code Assist, is available for all Google plans, and fits easily into both solo and team workflows.
Open Source and Customizable: Fully open-source, allowing you to tailor the tool to your own needs or contribute improvements to the wider developer community.
What This Means for You:
Gemini CLI lets you skip the back-and-forth between browser and terminalāyou can ask for code help, automate chores, or look up docs without breaking your workflow. If youāre often switching tools or spending time on routine setup, this could save you a lot of hassle.
With generous free limits and open-source flexibility, itās easy to try out or tailor to your needsāwhether youāre coding solo, working in a team, or just exploring whatās possible with AI.
š¤š» Anthropic Explores Claudeās Role in Emotional Support

Anthropic has conducted in-depth research to understand how users interact with Claude for emotional support, companionship, and personal adviceāeven though the AI was not originally designed for these purposes.
Key Highlights:
Rare, but Real: Only about 3% of conversations with Claude are for emotional or personal support, and requests for companionship or roleplay are even rarer.
Wide Range of Topics: Users talk to Claude about career worries, relationships, loneliness, big life questions, and sometimes for guidance on managing mental health challenges.
Positive Impact: Many users feel better after these chats, with conversations often shifting to a more positive tone, especially in advice or coaching scenarios.
Prioritizing Safety: Claude doesnāt give risky or harmful advice and will gently push back if a conversation could impact someoneās well-being.
Not a Therapist: Anthropic is clear that Claude isnāt meant to replace professional mental health careāthis research is about understanding user needs and keeping things safe.
What This Means for You:
Thereās growing buzz about people forming deep connections with AI, but for Anthropicās Claude, these moments are still rare. Most users are focused on getting things doneānot seeking friendship or advice. Still, this data may only reflect one side of the story, as different platforms draw different users.
As more mainstream and entertainment-focused tools emerge, emotional and social uses of AI may gradually grow. For builders and users alike, itās a good time to observe how expectations shiftāand consider how AI will fit into everyday life.
š OpenAI API Major Upgrade: Deep Research, Live Web Search, Webhooks & Logprobs
Launching today in the OpenAI API:
- Deep Research in the Responses API!
- Web search with o3/o4-mini!
- Webhooks!
- Logprobs in the Responses API!Please try to enjoy all of these ships equally š
ā Steve āļø (@stevendcoffey)
5:23 PM ⢠Jun 26, 2025
OpenAI has just introduced a powerful set of API enhancements, aimed at giving developers more advanced tools for automated research, real-time data, and deeper insight into model decision-making.
Key Highlights:
Deep Research API: Enables multi-step research workflowsāmodel plans sub-questions, executes web searches and code, and returns structured, citation-rich answers via the responses endpoint using o3-deep-research-2025-06-26 or lightweight o4-mini-deep-research models.
Pricing Tiered: o3-deep-research is priced at $10 per million input tokens and $40 per million output tokens; o4-mini-deep-research costs $2/$8 respectively.
Real-Time Web Search: Live internet data integration now supported via o3 and o4-mini, enabling apps to fetch and synthesize current information.
Webhooks Added: Developers can trigger async jobs and receive callbacks when tasks completeāideal for handling long-running research or workflow events.
Logprobs for Transparency: API now returns token-level log-probabilities alongside responses, offering deeper insight into model confidence and reasoning.
What This Means for you:
Developers can now embed true research intelligence into appsāautomating tasks like data gathering, analysis, and documentation with structured, verifiable outputs. Use web_search for fresh data, logprobs for quality control, and webhooks to orchestrate workflowsāoffloading complexity while keeping insight at your fingertips.
Start by automating one routine research taskālike fetching weekly market summaries or building citation-rich briefsāand iterate from there.
āļø Trending Bytes
TikTok: Gen Z
Instagram: Millennials
X: Chaos
ChatGPT: All of the above + your resume + your grocery list š§ š
our engineering and compute teams do incredible work to rapidly scale to meet customer demand for chatgpt.
a lot of blood sweat and tears go into this, and they make it look relatively easy.
i have never seen a team handle a 2.5 year sprint with such grace!
ā Sam Altman (@sama)
2:12 PM ⢠Jun 24, 2025
š ļø Quick Start Guide
Get started with Gemini CLI ā your AI assistant in the terminal
Step 1
Install Node.js (v18 or later) if you havenāt already.
Step 2
Run Gemini CLI directly from your terminal:
npx google-gemini/gemini-cli
Step 3
Log in with your Google account when prompted.
Step 4
Open an existing project folder or create a new one.
Step 5
Launch the CLI:
gemini
You can now use natural language to generate code, automate workflows, or explore docsāall right from your terminal.
š” AI Model Spotlight
Model Name | Parent Company | Release Date | Key Highlights |
---|---|---|---|
June 26, 2025 | CLI agent, 60 RPM limit,Text/code/tasks, Open-sourced via Max Text | ||
June 26, 2025 | Text, audio, image, Works offline, Privacy-friendly, Android/edge-ready | ||
HeyGen | June 26, 2025 | Multi-avatar control , Lip-sync AI, Auto-video scripts | |
Black Forest Labs | June 26, 2025 | Kontext Dev Tools ,Model registry, Open weights | |
Deepmind | June 25, 2025 | Gene mapping, Biotech R&D, Drug discovery use | |
Anthropic | June 25, 2025 | In-chat coding, Live editor, Real app deployment | |
ElevenLabs | June 23, 2025 | Fast voice cloning, Works across agents, Conversational memory | |
Mistral AI | June 20, 2025 | Better prompts, Less repetition, Strong functionācalling | |
June 17, 2025 | Stable GA, Deep Think mode-Thinking budgets | ||
June 17, 2025 | GA release- Cost-effective- High throughput | ||
June 17, 2025 | Preview launch- Fastest 2.5- Control thinking budget | ||
Krutrim | June 12, 2025 | Task-smart, Auto-execution | |
OpenAI | June 10, 2025 | Science-focused, Token-based | |
Mistral AI | June 10, 2025 | COT-ready, Compact models | |
Apple | June 9, 2025 | On-device 3B, Visual overhaul | |
Rednote (Xiaohongshu) | June 6, 2025 | Open-source model- Codingācapable- Hugging Face distribution | |
June 5, 2025 | Elo: +24 / +35- Deep Think- Thinking budgets | ||
Mistral AI | June 4, 2025 | Vibe coding- Onāprem/cloud- Multi-model integrate | |
OpenAI | June 3, 2025 | Internetāenabled- Dependency installs- Search & tests |
š Feedback
How would you describe your experience with this edition of the newsletter? |
Thatās it for nowākeep pushing AI forward! š
You received this email because you are subscribed to Outlier.ai. The content of this email is for informational purposes only and may not be reproduced or distributed without written permission. AI research is rapidly evolving, and while we strive for accuracy, we encourage readers to verify details from official sources.
Please note that all emails exchanged with Outlier.ai may be subject to monitoring for compliance with company policies. Information contained in this email is confidential and intended solely for the recipient. No legally binding commitments are created by this email.
All trademarks used in this email are the property of their respective owners. You are receiving this email because you have authorized Outlier.ai to send you updates. For more details, visit the Outlier.ai website. Terms & Conditions apply.