In partnership with
What if AI took over your job hunt tonight?
Meet AIApply.
Your AI Career Agent for modern job searching.
AIApply automatically:
Finds the best opportunities
Matches jobs to your profile
Customizes every application
Applies on your behalf 24/7
Saves hours every single week
Stop manually filling out forms.
Stop rewriting the same application over and over.
Whether you’re actively job hunting or casually exploring opportunities, AIApply keeps your applications moving around the clock.
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
Meta's AI Types Your Thoughts at 61% Accuracy, No Surgery Needed
TLDR: Meta built an AI that reads your brain activity and types out the words you're trying to write, and it works without any surgery.
The Story:
Meta showed off Brain2Qwerty v2 this week, an AI tool that turns brain signals into typed text. You wear a helmet-like scanner called an MEG machine, and the AI figures out the sentences you're trying to type. Meta trained it on about 22,000 sentences from nine people, each recorded for 10 hours while they typed. The system got the right word 61% of the time, a big jump from the 8% that older non-surgery methods managed. For the best person in the test, it hit 78%, with more than half the sentences off by one word or less. Meta is also giving away the training code so other scientists can build on it.
Its Significance:
Millions of people can't speak or move because of brain injuries, and until now the tools that let them type with their minds needed risky surgery to put electrodes inside the brain. This new approach skips that completely. It's still a research project, and the scanner costs millions of dollars and fills a room, so it's not coming home soon. But regular computers once filled rooms too, and now a more powerful one fits in your pocket, so brain scanners could shrink the same way. The flip side is worth sitting with: a machine that reads the words forming in your head is amazing for someone who lost their voice, but it raises real a whole bunch of concerns about the future of privacy and security, along and who gets to see your thoughts.
QUICK TAKES
The story: A UK law firm asked an AI chatbot to write a will for a made-up client, then had a probate lawyer grade it. The will read well, but it skipped key questions a real lawyer would ask, like taxes, trusts, or whether anyone should be left out.
Your takeaway: AI can copy the look of a legal document, but it doesn't know what you forgot to tell it. For big legal stuff, the questions matter as much as the writing.
The story: A Chinese company unveiled an AI that hunts for software security holes, calling it China's answer to Anthropic's restricted Mythos model. About a week and a half earlier, a different Chinese lab released a similar tool as free, open code that anyone can download.
Your takeaway: The US locked its most powerful cyber AI behind export controls. China is taking advantage by announcing its own and handing out a comparable version for free, so the wall the US built may not hold much back if China already has its own version. Anthropic has even accused the Chinese company Alibaba of using distillation attacks (training a copycat AI by feeding it thousands of answers from someone else's model, then learning to imitate it) to clone its models.
The story: A new report from Ramp and Revelio Labs looked at nearly 22,000 companies. The ones spending the most on AI, about $30 per worker each month, grew their headcount by 10.2%, and entry-level jobs at those firms rose 12%.
Your takeaway: The scary story is that AI wipes out junior jobs. This data points the other way, at least for firms that invest heavily. Companies that just buy a subscription and stop don't see those gains, so how much you commit seems to matter.
TOOLS ON OUR RADAR
📓 Anytype Free and Open Source: A beautiful private workspace that helps you build a digital second brain to organize your daily journals web clippings and creative ideas without relying on cloud servers.
🎵 Nuclear Free and Open Source: A desktop music streaming application that pulls audio from free sources across the web giving you a massive library of songs without any monthly subscription fees.
🎮 PPSSPP Free and Open Source: A brilliant emulator that allows you to play classic handheld console games in stunning high definition right on your modern computer or smartphone.
🌈 OpenRGB Free and Open Source: A wonderful customization tool that allows you to control the colorful lights on all of your computer hardware accessories from one single universal application.
TRENDING
Microsoft Builds a Bouncer to Keep Bots Out of Teams Meetings — Microsoft is rolling out a tool that spots third-party AI bots trying to sneak into Teams calls and makes a human approve them in the lobby first, so note-taking bots can't quietly record private meetings.
OpenAI Teases New Hardware for Codex — OpenAI is launching a small keyboard-like device called Codex Micro on July 15, built with Work Louder, giving coders physical buttons for their AI shortcuts. It's not the rumored AI phone with Jony Ive.
California Strikes Deal With Anthropic to Bring Claude AI to State Agencies — California signed a first-of-its-kind deal giving every state agency, city, and county access to Claude at a 50% discount, plus free training. The DMV and the state's Medicaid agency are already using it.
Gemini 3.5 Pro Cleared for July Launch as Fable 5 Nears Return, GPT-5.6 Stays Locked — Google's Gemini 3.5 Pro is on track for July as the only major new model with no government access limits, while Anthropic's Fable 5 eyes a comeback and OpenAI's GPT-5.6 stays locked to about 20 approved partners.
Gemini Can Now Take Notes for You in Google Meet — Google's "Take notes for me" now works for AI Pro and Ultra subscribers. It writes notes during your call, pulls out action items, saves everything to a Google Doc, and emails you a recap after.
AI Becomes an Early Step in the Homebuying Journey — More buyers now start house hunting by asking AI about budgets and neighborhoods before they ever call an agent. One study found 53% of would-be buyers would be comfortable buying a home with no human involved at all.
Cursor Brings Its AI Coding Agents to Mobile With a New App — Cursor launched an iPhone app that lets developers start and check on AI coding agents from their phone. Anthropic's Claude Code lead says most of his coding now happens on his phone.
TRY THIS PROMPT (copy and paste into Claude, ChatGPT, or Gemini)
🁢 Describe your goals. Find the one habit with the most leverage, the keystone that quietly tips all the others over.
Build a single-file HTML app with vanilla HTML/CSS/JS. The Keystone Habit Finder — describe your goals, get the one habit that triggers the most downstream change. Persist to localStorage key 'keystone_habit_v1'.
Aesthetic: warm dark (#0f1015), amber/gold (#f0b45a) primary with top radial glow. Fraunces serif (italic for prompts and the keystone name) for headings, Inter sans for body, JetBrains Mono for labels. Sage green (#78c8a0) for the anchor card, soft red (#e6786e) for the trap card. A vertical "domino chain" with numbered circles connected by a gradient line.
Form: goals textarea (encourage honesty/specificity), optional already-tried textarea, life-situation dropdown (busy / lots of free time / in a rut / chaotic), daily-energy dropdown (5 min / 15-20 min / 30+ min).
System instructions: behavior-change coach specializing in keystone habits (Duhigg concept) — the ONE habit that cascades into other positive changes. Identify the highest-leverage habit, not the obvious or most ambitious one; think in second/third-order effects; be realistic about life and energy; make the starting version absurdly small to guarantee a win; avoid encouraging unhealthy extremes. Return raw JSON: keystone (clear simple habit), why_keystone (2-3 sentences of leverage logic), dominoes (3-4 items: effect + detail on the mechanism), tiny_version (absurdly small starter), tiny_text (2-3 sentences, **bold** key instruction), anchor (existing routine to habit-stack onto), trap (likely self-sabotage + sidestep), runner_up (habit + why).
Render: centered gradient "keystone habit" hero card (label + big italic serif name + why). Domino-chain card with numbered circles connected by a vertical gradient line, each showing effect (bold) + italic detail. Amber "start absurdly small" card with the tiny version + bolded instruction. Two-column split: green "anchor it to" and red "trap to avoid." Runner-up card. Copy plan + archive of past keystones.What this does: Tell it what you want to improve, what you've already tried, your life situation, and how much daily energy you can spare. Instead of handing you a list of ten habits, it identifies the single keystone habit with the most downstream leverage, the one that makes the others easier almost automatically. You see why it's the keystone, a visual domino chain of the second and third-order changes it triggers, an absurdly small starting version so you can't fail the first day, an existing routine to anchor it to, the most likely way you'll sabotage it, and a runner-up if the first doesn't fit. Saves to localStorage.
What this looks like:
WHERE WE STAND(based on today’s news)
✅ AI Can Now: Turn brain activity into typed sentences at 61% word accuracy with no surgery, using a head scanner instead of implants.
❌ Still Can't: Do that outside a lab, since the brain scanner costs millions of dollars and fills a room.
✅ AI Can Now: Listen to a live meeting and write up a summary with action items on its own.
❌ Still Can't: Ask the follow-up questions a human expert would, like the missing details that make a legal will hold up.
FROM THE WEB
Free, open source AI models that you can download to your computer and use offline are nowhere near as good as the subscription models, but they are getting good enough for lots of basic tasks.
RECOMMENDED LISTENING/READING/WATCHING
A Memory Called Empire by Arkady Martine - Book
An ambassador from a small mining station arrives at the capital of a sprawling empire that has just absorbed her predecessor's body, but she's been given a malfunctioning version of her culture's "imago" tech, which is supposed to let her carry her predecessor's memories and skills directly inside her own head. Martine, a Byzantine historian, won the 2020 Hugo for this.
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.
Moda is the AI design agent with taste
Moda is an AI design product where you prompt what you need, get a complete on-brand design, and edit every element on a full canvas.
Our viral launch hit 4.4M views in days, tens of thousands signed up, and executives at major finance and tech companies now use it.