In partnership with
Attio is the AI CRM for high-growth teams.
Connect your email, calls, product data and more, and Attio instantly builds your CRM with enriched data and complete context. Whether you’re running product-led growth or enterprise sales, Attio adapts to your unique GTM motion.
Then Ask Attio to plan your next move.
Run deep web research on prospects. Update your pipeline as you work. Find customers and draft outreach emails. Powered by Universal Context, Attio's intelligence layer, Attio searches, updates, and creates across your data to accelerate your workflow.
Ask more from your CRM.
Beginners in AI
Good morning and thank you for joining us again!
Welcome to this daily edition of Beginners in AI, where we explore the latest trends, tools, and news in the world of AI and the tech that surrounds it. Like all editions, this is human curated and edited, and published with the intention of making AI news and technology more accessible to everyone.
THE FRONT PAGE
AI Lawyer Are Getting Good: Claude Now Scores 91% on Big Law's Toughest Test
TLDR: Anthropic released more than 20 new legal tools for Claude that connect to apps like DocuSign and Thomson Reuters, scoring 91% on Big Law's toughest AI test, while also opening free access for legal aid groups serving the 80% of people who face court without a lawyer.
The Story:
Anthropic just made a big push into the legal world. On Tuesday, the company rolled out 12 new Claude plug-ins for lawyers (covering commercial law, employment, privacy, M&A work, and even bar exam prep) plus 20 connectors that link Claude to tools law firms already use. That means a lawyer can now ask Claude to pull a contract from DocuSign, check case law on Westlaw via Thomson Reuters, or search files in Box without leaving the chat window. Anthropic's latest model, Claude Opus 4.7, scored 90.9% on Harvey's Big Law Bench, the test the legal industry watches most closely. Big-name firms including Freshfields, Quinn Emanuel, and Holland & Knight are already using it on live cases.
Its Significance:
Lawyers cost a lot of money. So when Anthropic says Claude can handle vendor contract reviews, deposition prep, and document search, that's work clients used to pay billable hours for. The bigger story might be on the other end. About 80% of people in civil court don't have a lawyer at all, and Anthropic is giving free Claude access to legal aid groups like Free Law Project and Courtroom5. So whether you're paying $800 an hour or fighting an eviction notice with no help, AI just walked into your case. This also puts Harvey, the AI law startup valued at $11 billion in March, in a strange spot. Harvey is built on Claude's underlying models, and now Anthropic is selling directly to the same law firms. The whole legal AI middle layer just got squeezed. Another thing worth knowing: lawyers keep getting caught filing court briefs with fake citations the AI made up, and judges have been handing out sanctions. The tools are here. The mistakes are still here too.
QUICK TAKES
The story: Google introduced Googlebook on Tuesday, a premium laptop line with Gemini baked into the operating system. Dell, HP, Lenovo, Acer, and Asus are all making models, launching this fall. The laptops have a "Magic Pointer" cursor that uses AI to suggest actions based on what you're hovering over, and they sync with Android phones so you can use phone apps right on the laptop.
Your takeaway: Google is making a real push to challenge MacBooks and Windows PCs for the high end of the market, not the cheap Chromebook space. No pricing yet. If you're shopping for a laptop this summer, it's worth waiting until the fall to see what the first Googlebooks actually deliver or if the features are all something that you can use on your regular computer without buying a brand new one.
The story: The NYT sent freelancers a strict reminder that they can't use ChatGPT, Claude, Gemini, Perplexity, or any AI image tool to "create, draft, guide, clean up, edit, improve, or rephrase" their work. The email came after a Times bureau chief published an article with a fake quote that turned out to be AI-generated, plus a book reviewer got fired in March for AI-written content.
Your takeaway: If you're writing for any publication right now, assume the rules are tightening. Brainstorming with AI is okay at the Times. Touching the actual words with it isn't. The newspaper says staff reporters have separate (more lenient) rules, which already has some critics asking why freelancers carry the whole risk.
The story: RSL Media, a new nonprofit backed by Cate Blanchett, George Clooney, Viola Davis, Tom Hanks, and Meryl Streep, lets anyone (not just celebrities) declare how AI systems can use their work, name, likeness, or voice. The free public registry opens in June. You pick one of three signals (allowed, allowed with terms, or prohibited) and AI companies that respect the standard can read it like a traffic light.
Your takeaway: This isn't a law, so AI companies can still ignore it. But it's the first time normal people will have a machine-readable way to say "no" to having their face or writing scraped. Worth signing up when the registry opens, especially if you post a lot of writing, art, photos, or video online.
TOOLS ON OUR RADAR
🍿 Jellyfin Free and Open Source: A phenomenal media server software that transforms your computer into a personal streaming service beautifully organizing your entire collection of digital movies and television shows.
📝 AppFlowy Free and Open Source: A highly customizable digital workspace that serves as a fantastic alternative to Notion allowing you to build wikis organize tasks and manage personal databases completely offline.
🖼️ Immich Free and Open Source: A blazing fast photo and video backup solution that rivals Google Photos allowing you to automatically sync your memories straight from your mobile phone to your own private server.
🍲 Mealie Free and Open Source: A visually pleasing recipe manager that automatically extracts ingredients and cooking instructions from your favorite food blogs allowing you to easily plan out your weekly dinners.
TRENDING
Big Study Finds AI Layoffs Aren't Actually Paying Off - Gartner surveyed 350 executives at companies bringing in $1 billion or more a year and found that 80% had cut staff to invest in AI. The kicker? Companies that kept their workers and gave them AI tools saw the same financial gains as the ones that did layoffs, suggesting the cuts didn't actually help.
Hackers Slip Malware Into a Mistral AI Software Package - Microsoft caught attackers planting credential-stealing code inside a Mistral AI Python package on PyPI, hitting Linux developer machines. The malware was part of a bigger campaign called "Mini Shai-Hulud" that compromised over 170 packages with combined weekly downloads of 518 million. Mistral says its own systems weren't breached.
Fake OpenAI Repo Hit #1 on Hugging Face, Stole Passwords for 18 Hours - A fake repository copying OpenAI's Privacy Filter model racked up 244,000 downloads on Hugging Face before getting pulled. The malware stole browser passwords, Discord tokens, crypto wallet keys, and SSH credentials. Most of the 667 "likes" came from bot accounts gaming the trending algorithm.
Android Phones Are Getting Gemini Intelligence This Summer - Google's pushing AI deeper into Android with multi-step task automation, smarter autofill, voice-to-text cleanup, and custom widgets you can build by describing them. Samsung Galaxy S26 and Google Pixel 10 get it first.
Google DeepMind Wants to Reinvent the Mouse Cursor - DeepMind published research and demos for an AI-powered pointer that understands what you're hovering over. Point at a recipe and say "double these ingredients." Point at a chart and ask for a pie chart version. The feature is rolling out as "Magic Pointer" on Googlebook and inside Chrome.
Anthropic Is About to Pass OpenAI in Business Customers, Per Ramp - Anthropic's share of paying business customers jumped from 24.4% to 30.6% in a single month, closing the gap with OpenAI (35.2%) to just 4.6 points. Among VC-backed companies, Anthropic already leads 66% to 59%. Anthropic's revenue run-rate hit $30 billion, up from $9 billion at the end of 2025.
New Startup Adaption Launches AutoScientist, an AI That Helps Train Other AIs - Sara Hooker (former VP of AI research at Cohere) launched a tool that automates how AI models learn new skills, claiming it more than doubled win rates across different models. The bigger claim is that this could let companies outside the giant AI labs train frontier-level models.
Medicare Just Created a Payment Path Specifically for AI Healthcare Agents - A new 10-year Medicare program called ACCESS goes live July 5, paying companies to use AI agents for managing patients with conditions like diabetes, high blood pressure, and depression. Payment is tied to actual health results, not just activities. 150 participants are in the first cohort.
TRY THIS PROMPT (copy and paste into Claude, ChatGPT, or Gemini)
🎤 Name any field. Get the 10 questions an expert would ask that a beginner would never think of.
Build a single-file HTML app using vanilla HTML, CSS, JS, and one API call. Create The Expert Interviewer — a tool that generates the 10 questions an expert would ask in any field. Use localStorage key 'expert_interviewer_v1'.
Aesthetic: near-black (#0d0e14) with a subtle dot-grid overlay, radial vignette. Playfair Display italic 900 for headings, Inter for UI, Courier Prime monospace for labels. Purple/violet (#a09ad0) accent. Question numbers in large italic serif.
Form: field name input, question focus dropdown (6 options: general fluency / hiring / decision-making / spotting fraud / learning from scratch / good vs great), current level dropdown, optional context input.
Call the API instructing it to generate questions a beginner would never think to ask — specific, non-obvious, revealing. Return raw JSON: field_name, expert_edge (2 sentences on what separates expert from knowledgeable amateur), questions array (number, category, question, why_it_matters in 1 sentence), deepest_cut (the single best question), red_flag (the answer that immediately signals a novice or fraud).
Render: a field header with italic Playfair Display field name and focus meta, an expert edge section explaining what separates real experts, the 10 question cards each with large italic number, category label in monospace, the question in large serif, and why-it-matters in small Inter below, a two-column bottom row for deepest cut and red flag. Save sessions to localStorage with field name and date. Make it work in a single HTML file.What this does: Type any field: venture capital, emergency medicine, wine, machine learning, architecture, supply chain. Choose a focus (general fluency, hiring someone, spotting a fraud, learning from scratch) and your current level. The Expert Interviewer generates 10 non-obvious questions that expose real depth, each tagged by category (Mental Models / Trade-offs / Failure Modes / Edge Cases / Judgment) with one sentence on what a strong answer actually reveals. A closing section surfaces the single question that cuts deepest and the most common answer that sounds smart but immediately signals a novice. Every session saves to localStorage so you can build a personal library of expert lenses across fields.
What this looks like:
Global HR shouldn't require five tools per country
Your company going global shouldn’t mean endless headaches. Deel’s free guide shows you how to unify payroll, onboarding, and compliance across every country you operate in. No more juggling separate systems for the US, Europe, and APAC. No more Slack messages filling gaps. Just one consolidated approach that scales.
WHERE WE STAND(based on today’s news)
✅ AI Can Now: Pull a contract from DocuSign and check case law in Westlaw inside a single Claude chat
❌ Still Can't: Stop making up fake case citations that get lawyers sanctioned in court
✅ AI Can Now: Manage chronic care patients between visits with Medicare paying for the work
❌ Still Can't: Write a reliable freelance article without possibly getting the freelancer fired via AI language patterns
FROM THE WEB
RECOMMENDED LISTENING/READING/WATCHING
A Michael Winterbottom film starring Tim Robbins and Samantha Morton, set in a near-future where movement and reproduction are controlled by predictive algorithms that read your genetic data and decide whether you're allowed to be where you are or with whom. Quietly one of the most accurate predictions of algorithmic governance ever filmed.
Thank you for reading. We’re all beginners in something. With that in mind, your questions and feedback are always welcome and I read every single email!
-James
By the way, this is the link if you liked the content and want to share with a friend.
Some * designated product links may be affiliate or referral links. As an Amazon Associate, I earn from qualifying purchases. This helps support the newsletter at no extra cost to you and Amazon makes a tiny hair less.
