Everything I'd send you, in one place

When AI Builds Itself (Anthropic)

Anthropic Institute essay on progress toward recursive self-improvement in AI and its implications.

Review GitHub Pull Requests with Codex

OpenAI's official guide to configuring Codex code and security reviews for GitHub pull requests, including automatic reviews and custom repository instructions.

NVIDIA CUDA Toolkit

The official free download and documentation for CUDA, Nvidia's platform for running AI and scientific computing on its graphics chips. Includes compilers, libraries and learning guides.

Why I recommend it: The toolkit itself is free, but it only runs on Nvidia's own chips — learning CUDA ties your skills to one company's hardware. Nvidia dominates AI chips, so that trade-off is real but worth knowing about.

Claude Corps — Anthropic's AI Fellowship for Nonprofits

A fellowship from Anthropic that places early-career people with mission-driven nonprofits to put AI to practical use. You can apply to be a fellow, or a nonprofit can apply to host one. The page explains how the programme works and how to apply.

Why I recommend it: This is Anthropic's own program page, so it presents the fellowship in the best light. Check the current application windows, pay and eligibility on the page before applying — those details change from cohort to cohort.

Building a C compiler with a team of parallel Claudes (Anthropic)

Anthropic engineering write-up on running many Claude agents in parallel to build a working C compiler, including testing against GCC.

Why I recommend it: Anthropic describing its own product's success. The details are useful, but the results have not been independently reproduced.

Emotion concepts in a large language model

Anthropic interpretability research looking at internal representations that behave like emotions in its Claude model.

Why I recommend it: Written by the company that builds the model. "Emotion concepts" are patterns in the model, not proof it feels anything.

Claude Discovers a Novel Enzyme System

Anthropic reports early results from its life-sciences lab, where Claude agents identified an enzyme system whose function is still unknown.

Why I recommend it: Anthropic describing its own product. Early results; wait for peer review before treating it as a confirmed discovery.

Trusted Contacts in ChatGPT

OpenAI help article on Trusted Contact: an optional adult (18+) feature that may notify one person you choose if automated systems and trained reviewers detect a serious suicide-related safety concern.

Why I recommend it: Worth reading before you turn it on: it involves human reviewers reading flagged conversations and sharing an alert with someone else. OpenAI says it is not an emergency service. Not available in Business, Enterprise, or Edu workspaces.

Introducing MentalHealthBench

OpenAI's benchmark of 1,215 realistic mental-health conversations, scored against rubrics written by 80+ licensed mental-health experts, covering everyday well-being through emergencies across ages and languages.

Why I recommend it: OpenAI built this benchmark and grades its own models on it, so treat the "steady progress" claim as a self-report until outside researchers replicate it. Useful for its honest list of weak spots: asking for context and judging urgency.

Crisis Helpline Support in ChatGPT

OpenAI help article explaining the localized crisis helplines ChatGPT surfaces (built with ThroughLine) and how to use a crisis line. In the US, call or text 988.

Why I recommend it: A plain guide to what a crisis line is and how to reach one. ChatGPT is not a crisis service; if you or someone you know is in danger, contact a helpline or emergency services directly.

Sycophancy in GPT-4o: What Happened and What We're Doing About It

OpenAI's April 29, 2025 explanation of why it rolled back a ChatGPT update that made the model overly flattering and agreeable. It says the update leaned too heavily on short-term thumbs-up feedback, and lists the fixes it planned.

Why I recommend it: A company explaining its own mistake, so read it as OpenAI's account, not an independent review. Useful for seeing why a chatbot that always agrees with you isn't a reliable advisor.

The Three Stats to Know This Academic Year (replay)

A 58-minute recorded webinar for college career-services staff and advisers, run by Jobscan with Jeremy Schifeling of The Job Insiders, on three labour-market numbers worth building this year's student advising around — including hiring data and H-1B sponsorship patterns.

Why I recommend it: Free to watch with no sign-up, recorded live on 22 September 2026. Be straight with yourself about what it is: a raw, unedited replay shared on a private Descript link by the host, which means it can be taken down at any time, and it is a webinar run by a company that sells resume-scanning software, so expect the product to appear. I could not read the full transcript, so I am not repeating the three figures here — watch it and check each number against its named source before you quote it to a student. Jeremy Schifeling also published a free H-1B sponsor-finder tool alongside the event.

OpenAI Advisory Group on Mathematics and AI

OpenAI is funding an independent advisory group of mathematicians, hosted at the Institute for Advanced Study, to advise on how AI-generated math results are shared.

Why I recommend it: Announced alongside a claim that an internal model solved 100+ open math problems in a month. The group is explicitly not allowed to slow the pace of research — read it as a communications channel, not a brake. Members include Terence Tao and Timothy Gowers.

Introducing Claude Opus 5.5 (Anthropic)

Anthropic's own announcement of Claude Opus 5.5, published 22 September 2026 — the first model in the Claude 5.5 family. The company's claims, in its own words: it performs at the level of Claude Fable 5.1 on most work and costs 40% less to run than Opus 5 on typical workloads; on Anthropic's automated behavioural audit, its most comprehensive internal alignment test, it scores the highest of any model the company has tested. The page also cites early-tester anecdotes, including one completing a 680,000-line code migration in under a day, and a result of succeeding 39 times out of 40 on a page-load optimisation task. It is Anthropic's first release since the company publicly called for pacing the frontier, and it was tested before release by external evaluators including METR and Frontier Design.

Why I recommend it: Read this as a primary source, not as a review. Every performance and cost figure on the page was produced by the company selling the model, on tests it designed — that does not make them false, it makes them unchecked by anyone with a reason to doubt them. The checkable part is the external pre-release testing by METR and Frontier Design, so that is the part to weigh. Note too that '40% less to run than Opus 5' compares Anthropic with its own older model and says nothing about a rival's price, and that the standout numbers come from hand-picked early testers rather than a measured success rate you can plan around. The announcement is free to read; using the model itself is not free beyond whatever the current Claude free tier allows.

OpenAI AI and Teen Development Research Grants

OpenAI is committing $5 million, with individual grants up to $1 million, to fund independent research into how generative AI affects young people aged 13-17, with a focus on social and emotional development. Topics include how teens actually use AI, developmental outcomes, the factors that shape those effects, and which safeguards and design choices work. Applications opened 8 September 2026 and close 6 October 2026, 11:59 PM PDT, reviewed on a rolling basis with decisions by 13 November 2026. Applicants must be 18 or older and affiliated with a research institution or have significant relevant experience; proposals are welcome from any country. Free to apply.

Why I recommend it: Relevant if you do research, teach, or work in youth services and have a study you cannot fund — the eligibility wording allows significant relevant experience as an alternative to an institutional affiliation, which is wider than most AI grants. Say the obvious thing plainly, though: OpenAI is funding research into the effects of its own category of product, and it chooses who gets the money. That does not make the findings wrong, but disclose the funder in anything you publish. Deadline 6 October 2026, and check the dates on OpenAI's own page before you rely on them.

Jev AI vs LLMs: When Should You Use a Decision Model Instead of a Chat Model?

A practical engineering comparison of decision models (Jev AI) and generative large language models: when to use each, how they can work together, and a five-dimension framework for choosing.

Why I recommend it: Community article on Hugging Face by sora-2, published 21 September 2026. It explains that Jev AI turns state into a choice, score or yes/no judgment inside a defined answer space, while generative LLMs handle open-ended writing, explanation and reasoning. Rule of thumb: if the answer space can be defined in advance and the result will be reused, ranked, routed or blocked by code, evaluate Jev AI first.

OpenAI Model Spec (2026-08-18)

OpenAI's published specification for how its models are supposed to behave: red-line principles, the chain of command between platform, developer and user instructions, content boundaries, and how conflicts should be resolved. Free to read in full, no sign-up.

From the site: The Model Spec specifies desired behavior for the models underlying OpenAI

Why I recommend it: Useful as a primary source when people argue about what an AI assistant "should" do — this is the maker's own stated rulebook, so read it as OpenAI's intent rather than an independent audit of actual behavior.

Mistral Le Chat

Mistral AI's assistant: web search with citations, document and image analysis, image generation and task agents. Free tier is generous; a paid tier with higher service guarantees is planned but not needed to use it.

Why I recommend it: Good second opinion when you want sourced answers with citations you can click through and check.

Claude

Anthropic's AI assistant. Strong at long writing, reading documents you paste in, analysis and coding help. Free tier with daily limits; paid plans lift them.

From the site: Claude is Anthropic

Why I recommend it: The one I reach for when the task is writing or thinking through a document. Check anything factual yourself — it can be confidently wrong.

Gemini

Google's AI assistant, wired into Gmail, Docs and the rest of Workspace, so it can work on files you already have. Free tier; paid plans add the larger models.

Why I recommend it: Worth it mainly if your work already lives in Google Docs and Gmail — that integration is the real advantage.

Model Misalignment Reporting Framework (OpenAI)

OpenAI's free framework for how misaligned model behaviour should be reported and categorised — what counts as misalignment, who reports it, and what happens next.

From the site: OpenAI shares a framework for tracking, investigating, and disclosing model misalignment, alongside six reports of unexpected or concerning model behavior.

Why I recommend it: Primary source on how a major lab defines and handles its own model failures — useful, but it is the lab grading itself.

OpenAI Alignment — research and releases

OpenAI's free alignment research hub, including reports documenting how its own models fail.

From the site: Research on aligning AI with human values and intent, and reports documenting model failures.

Why I recommend it: A lab publishing on its own safety work — valuable primary material, but not an independent audit.

Perplexity

An AI answer engine that searches the live web and cites its sources, with a free tier that covers everyday research.

Why I recommend it: Good for a first pass on a company, an industry, or a role — then click the citations and read the originals yourself.

Perplexity Research Residency

A residency at Perplexity for people who want to do applied AI research on a working product rather than in a lab.

Why I recommend it: A residency is the friendliest door into research — you are paid to learn on a real team instead of needing the credential first.

Claude Corps — Anthropic's AI Fellowship for Nonprofits

A fellowship from Anthropic that places early-career people with mission-driven nonprofits to put AI to practical use. You can apply to be a fellow, or a nonprofit can apply to host one. The page explains how the programme works and how to apply.

Why I recommend it: This is Anthropic's own program page, so it presents the fellowship in the best light. Check the current application windows, pay and eligibility on the page before applying — those details change from cohort to cohort.

Career Compass Guide (ChatGPT)

A ready-made ChatGPT assistant for career direction — talk through where you are, what you want next and the steps in between. Usable with a free ChatGPT account.

ElevenLabs Seedream 4.5 Image Generation

ElevenLabs' hosted access to the Seedream 4.5 image model, with a free tier for trying image generation for slides, social posts and simple brand assets.

Why I recommend it: Good enough for pitch decks and social graphics — check the free-tier limits before you plan a big batch.

OpenAI Realtime Voice Prompting Guide

Official OpenAI guide for designing prompts for Realtime voice models, including gpt-realtime-2 and gpt-realtime-1.5. Covers role definition, guardrails, tool delegation and iterative testing.

Why I recommend it: Start here if you are building voice agents or want cleaner spoken-AI interactions. The guide recommends starting minimal and adding rules only for behaviors that fail in testing.

Kickresume

Resume and cover letter builder with a large template library, an AI writer, and a free tier for basic documents.

Why I recommend it: Pick one clean template and reuse it. Switching designs every application wastes hours you should spend researching companies.

OpenAI Computer Use Documentation

Official documentation for ChatGPT computer use, explaining what the agent can do and its safety limits.

Why I recommend it: Read the limits section carefully before you hand an agent anything connected to your accounts.

Apify CLI for AI Agents

A walkthrough of using the Apify command line tool to let AI agents run web scraping and automation tasks, aimed at people building their own small automations.

Why I recommend it: This is for the tinkerers. If you have ever wanted a repeatable way to pull data for lead lists or market research, this is a concrete starting point rather than another think piece.

Jobscan LinkedIn Profile Optimization Scan

A free scan that compares your LinkedIn profile against a target job title and returns a score plus specific fixes for your headline, about section, skills, and keywords - the same matching logic Jobscan uses for resumes, applied to your profile.

Why I recommend it: Run it once, fix the three biggest gaps it names, and stop there. The score is a diagnostic, not a target to chase.

OpenAI: The AI Policy Window

OpenAI's policy essay on the current window for AI regulation and the tradeoffs shaping government decisions.

Why I recommend it: Read this as a company making its case, not a neutral source. Useful for understanding the argument you will be asked to react to at work.

Anthropic Economic Scenarios

Research exploring possible economic futures as AI capability advances, including labor market effects and policy questions.

Why I recommend it: Scenario planning is a career skill, not just a policy exercise. Read it and ask which future your current job depends on.

An Alien Mind

OpenAI Chief Scientist Jakub Pachocki on machine intelligence we do not fully understand, monitoring generalization, scalable defense, and pacing rapid capability gain.

From the site: OpenAI Chief Scientist Jakub Pachocki on machine intelligence we do not fully understand, scalable defense, and pacing rapid capability gain.

Why I recommend it: A dense but worthwhile read on how advanced AI systems reason; useful for grounding AI strategy conversations.

OpenAI: GPT-6 Astra

OpenAI's announcement of GPT-6 Astra, its most capable model, with reported results on computer use, browsing, software engineering, cybersecurity, and professional work.

Why I recommend it: Read the capability list as a job-task list. Whatever a model does well this year reshapes entry-level work the next.

Jobscan Resume Scanner App

Jobscan's scanning workspace: paste a resume and a job description to see keyword match rate and ATS formatting issues. Free scans monthly, paid for unlimited.

Why I recommend it: Use your free scans on the roles you actually want. Match the language in the posting, but never paste in skills you cannot back up in an interview.

LLM Course (Hugging Face)

Deep dive into transformer architectures, tokenization, datasets, NLP pipelines, and fine-tuning and deploying open-source models.

Why I recommend it: Take this after a beginner course. It is where you learn how these models actually work under the hood.

OpenAI Academy

Free lessons on core AI concepts, prompting workflows, building and configuring AI agents, and applied ChatGPT use cases.

Why I recommend it: Practical rather than theoretical. Good for turning ChatGPT from a toy into part of your daily workflow.

AI Agents Course (Hugging Face)

Free course on autonomous and tool-using agent architecture, with framework implementations in smolagents, LlamaIndex, and LangGraph, plus agentic RAG workflows.

Why I recommend it: Agents are what employers are hiring for right now. Build one small working agent and you will be ahead of most applicants.

NVIDIA training and certification catalog

Searchable catalog of NVIDIA courses and certifications in AI, data science, and accelerated computing, including free self-paced options.

Why I recommend it: Go here after the general AI courses, when you want a technical credential rather than a literacy one.

Privacy-Safe Personalization

Blog post on how marketers can personalize campaigns while still respecting user privacy.

Speechify

Text-to-speech app that converts written content into natural-sounding narrated audio.

Runway

AI-powered video generation and editing platform for creating and stylizing clips from text or images.

Hugging Face

Open-source AI community and platform hosting machine-learning models, datasets and tools. Free to browse, download and run models; paid plans only cover hosted compute.

OpenAI Agentic Commerce Protocol

OpenAI's developer documentation for its protocol letting AI agents complete purchases on a user's behalf.

Descript

AI-powered audio and video editing tool that lets users edit recordings by editing a text transcript.

Anthropic

AI safety company that builds the Claude family of AI models with a focus on responsible AI development.

Sustainability & Ethical AI

Hugging Face blog post discussing the environmental sustainability and ethics of AI systems.

Otter.ai

AI transcription tool that records and transcribes meetings, interviews, and voice notes in real time.

Why I recommend it: A great free tool for transcribing client calls or podcast interviews.

OpenAI News / Dev Day

OpenAI's official newsroom covering product announcements, model updates, and Dev Day news.

Synthesia

AI platform that produces videos narrated by realistic AI avatars from a simple text script.

Pika

AI video generation tool that turns text and image prompts into short animated video clips.

Voiceflow

Platform for designing and building AI chatbots and voice assistants without heavy coding.

Sora 2

OpenAI's generative AI model that creates realistic video and audio clips from text prompts.

Predis.ai

AI tool that generates social media posts, captions, and videos from a website URL or prompt.

ChatGPT

OpenAI's conversational AI assistant for writing, research, brainstorming and support tasks. Free tier with message and model limits; paid plans lift them.

Why I recommend it: The single most useful free tool for brainstorming marketing copy and business ideas.

Blaze AI

AI content-generation platform for creating marketing copy, social posts, and blog content.

OpenAI: Our decision on Cursor following its acquisition by SpaceX

OpenAI's official statement explaining why it is winding down the contract that supplied its models to Cursor (Anysphere) after SpaceX completed its $60B acquisition of the AI coding company in August 2026.

In plain terms: OpenAI says it will stop supplying its models to the AI coding tool Cursor after SpaceX bought the company, citing concerns about terms-of-service compliance. Cursor users may lose access to OpenAI models, so the practical takeaway is not to depend on a single AI tool or provider.

Why I recommend it: A clear-eyed lesson in platform risk: the tools you build your workflow on can lose access to the models that make them work. If you code, write, or job hunt with an AI tool, know which models sit underneath it and keep a backup you already know how to use.

NVIDIA Free Training Courses

NVIDIA's catalog filtered to its free self-paced courses covering AI, deep learning, generative AI, data science, accelerated computing, and CUDA — with certificates of competency on many tracks.

In plain terms: This website offers a collection of free, self-paced online classes in artificial intelligence, data science, and computing. You can take lessons to build new technical skills and earn certificates to share with employers.

From the site: Browse NVIDIA self-paced and instructor-led training, including free courses in AI, deep learning, generative AI, data science, and accelerated computing.

Why I recommend it: Start with a free intro course and put it on your resume under a "Continued Learning" section. Hiring managers notice vendor-name training, and NVIDIA carries weight in AI and data roles.

Teal Resume Builder

Free resume builder that scores your resume against a specific job description and suggests the keywords you're missing.

In plain terms: This free online tool helps you create and update your resume to match specific job postings. You can compare your draft against any job description to find missing keywords and download your finished document.

From the site: Create ATS-friendly, job-matched resumes in minutes. Works with any job description.

Teal AI Resume Builder

Tailors your resume to each job posting and tracks your applications in one place.

In plain terms: This online tool helps you create resumes and organize your job search in one place. You can tailor your resume keywords to match specific job descriptions and track all your applications from start to finish.

From the site: Create a professional resume in minutes with Teal. Create unlimited personalized resumes for every job you apply for.

Why I recommend it: The free tier is useful. Let it draft, then you edit for truth and voice.

Free ATS-Friendly Resume Templates (Jobscan)

Clean, ATS-friendly resume templates you can edit in your browser for free.

Why I recommend it: Pick the simplest one. Fancy columns confuse applicant tracking systems.