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
Elon's New Grok Is 60% Cheaper Than Claude, and a Year Behind
TLDR: Elon Musk's SpaceXAI released Grok 4.5, a coding-focused AI that costs less than half what Claude and ChatGPT charge, runs fast, and only scores a step below the top models by Musk's own account.
The Story:
SpaceXAI put Grok 4.5 out on July 8. It costs $2 for a million input tokens and $6 for a million output tokens, the small chunks of text an AI reads and writes. Claude Opus 4.8 runs $5 and $25, and OpenAI's new GPT 5.6 Sol runs $5 and $30, so Grok is the cheap one. Musk called it "roughly comparable to Opus 4.7, but much faster." That's a careful line, since Opus 4.7 was Anthropic's flagship before 4.8 took over, so he's really saying it matches last year's best Claude, not this year's. On a bug-fixing test it scored 53%, while Anthropic's top model Fable 5 hit 70%, so it's a step behind on raw skill. But it uses far fewer tokens per job and runs at 80 tokens per second, so teams can do more work for less money. It caps a busy year for Grok, which has added video generation, voice cloning, and a coding tool that goes up against Claude Code.
Its Significance:
Most people won't pick an AI by who tops a leaderboard. They'll pick by what they can afford. A cheaper, faster model gives small teams and solo builders a real option they didn't have last week, and if you write code or pay for AI by the token, Grok 4.5 could cut your bill without slowing you down. When one company drops its price this hard, the others usually follow, which is good news no matter which AI you end up using. And for the price, this is still a highly capable model for plenty of everyday tasks.
QUICK TAKES
The story: OpenAI posted a job for an investment banking expert, paying up to $205,000 plus equity, to teach its AI how bankers handle deals like mergers and fundraising. The goal is to set the "quality bar" so ChatGPT can do the grunt work junior analysts now spend 80-hour weeks on.
Your takeaway: This lines up with what Citadel CEO Ken Griffin told students at Stanford this spring. Griffin, who used to call AI overhyped, said work that once took his master's and PhD finance staff weeks or months is now getting done by AI in hours or days. He said watching it happen inside his own firm left him "fairly depressed." When a job's routine parts can be handled by a machine, the people who stay are the ones who can check the AI's work and catch its mistakes.
The story: A free tool called Claude Wrapped lets you upload your Claude chat export and see a Spotify-Wrapped-style rundown of your stats, your writing habits, and your collaboration style. It's made by an independent developer, not Anthropic, and it runs entirely in your browser so your data never leaves your device.
Your takeaway: It's a fun way to notice how you actually use AI, like which topics you lean on it for and when. Just know the difference: this is a third-party project. Anthropic has its own separate "reflect" feature built into Claude. If you try the third-party one, the private, on-device setup is a nice touch. Tools like this are becoming more prolific and it's just as easy to make your own, more private version.
The story: Meta rolled out Muse Spark 1.1, which its AI chief Alexandr Wang calls the company's "strongest model for agentic and coding work yet." New accounts get $20 in free credits, then pay $1.25 per million input tokens and $4.25 per million output. Wang called the pricing "very aggressive" next to Anthropic and OpenAI.
Your takeaway: Meta was late to this and it shows, but low prices pull in developers fast. More players fighting over coding tools means better deals and faster updates for anyone who builds with AI. Watch whether Meta can keep shipping updates instead of going quiet after the launch.
TOOLS ON OUR RADAR
🎙️ AudioPen Freemium: A web and mobile application that takes your unstructured voice ramblings and uses AI to clean them up into perfectly formatted text, meeting notes, or blog posts in seconds.
🐧 LocalSend Free and Open Source: A cross platform file sharing application that enables secure communication between your devices without needing the internet. It allows you to instantly drop files across Mac, Windows, Android, and iOS over your local network. (Alternative to Apple AirDrop)
📊 Plus AI Freemium: An extension that lives directly inside Google Slides and PowerPoint. It allows you to generate complete, professional slide decks from a single prompt while automatically matching your specific company branding and formatting.
🛋️ RoomGPT Freemium: A user friendly AI room designer that lets you upload a photo of your current space, select a design style, and instantly generate realistic visual concepts of what your room could look like with a completely new aesthetic.
TRENDING
OpenAI Made ChatGPT Better at Small Talk - A new voice model called GPT-Live can say "mhmm" or "yeah" while you talk, jump into quick back-and-forth, or stay quiet when you need to think. It listens and responds at the same time instead of waiting for you to finish, so chatting feels more like a real conversation.
OpenAI Published Its Rules for Working With the Government - OpenAI laid out its national security principles, promising no mass domestic surveillance, no controlling autonomous weapons, and no high-stakes automated decisions. It's expanding cyber and biosecurity work with the U.S. and allies like Canada, Japan, and South Korea.
Google Photos Can Now Restyle Your Videos in Seconds - A new Video Remix tool, powered by Google's Gemini Omni model, lets you brighten a dark clip, swap the background, or add painterly looks like watercolor and oil paint using ready-made templates. It's rolling out to Google AI Plus, Pro, and Ultra subscribers in select countries.
News Outlets Ask a Judge to Punish OpenAI in a Copyright Fight - The New York Times and other papers asked a Manhattan court to sanction OpenAI, saying it lied about being unable to search its systems for their articles and deleted billions of ChatGPT chats. They want attorney fees and a ruling that OpenAI misused their work. The case started back in 2023.
Meta Launched Its Own AI Image Maker - Meta released Muse Image, its first in-house image model, free inside the Meta AI app, Instagram Stories, and WhatsApp. It can even pull public Instagram photos into your creations, which raised privacy concerns, though you can opt out in settings.
A Brown Professor Says Most of His Class Used AI to Cheat - Economics professor Roberto Serrano gave a take-home midterm and watched the class average hit 96, way above the usual 65 to 80. When he switched the final to an in-person test, the average dropped to 48. He believes at least 50 students cheated with AI.
TRY THIS PROMPT (copy and paste into Claude, ChatGPT, or Gemini)
⚙️ Describe the task you keep rebuilding from scratch. Get a step-by-step system so you stop reinventing it every time.
Build a single-file HTML app with vanilla HTML/CSS/JS. The Systemize-It Coach — describe a recurring task, get a step-by-step reusable system. Persist to localStorage key 'systemize_coach_v1'.
Aesthetic: dark navy-black (#0b0f16), blueprint blue (#4a9eff) primary with top radial glow. Sora sans for headings, Lora serif italic for overview/notes, JetBrains Mono for labels and the template block. Amber (#e6965a) for the pitfalls card, green (#78c896) accent for the time-saved card.
Form: recurring-task textarea, frequency dropdown (few times a week / weekly / monthly / a few times a year), who's-involved dropdown (just me / small team / family-household / clients-vendors), optional pain-points textarea (what trips you up).
System instructions: systems/process coach turning a redone task into a reusable system like an SOP; fixed checklist for the unchanging parts, decision points (if/then) for the parts that branch, an optional copy-paste template for anything involving writing/communication, common failure points each with a fix; address stated pain points directly; practical not bureaucratic; no em dashes. Return raw JSON: system_name, overview (2-3 sentences), meta_chips (3-4 tags), checklist (4-7 items: text + optional note), decision_points (2-4 if/then pairs, empty array if task doesn't branch), template (optional plain-text template with [placeholders], empty string if n/a), pitfalls (2-4 items: failure + fix in one sentence), time_estimate (1 realistic sentence).
Render: gradient system-name hero card (overview + meta chips). Numbered checklist card with boxed step numbers + optional italic notes. Blue "decision points" card with IF/THEN pairs (only shown if array non-empty). Dark mono "reusable template" code block with its own copy button (only shown if template present). Amber "common failure points" card with ⚠ items. Green time-saved card. Copy system + archive of past systems by name.What this does: Describe a task you redo constantly, how often it comes up, who's involved, and where it usually goes wrong. It builds a named, reusable system: a fixed checklist of the steps that never change, decision points for the parts that branch (if this, then that), an optional copy-paste template if the task involves writing or communicating something, and the common failure points with a fix for each, addressing your stated pain points directly. A closing line names roughly how much time or hassle the system saves once it's running. Everything saves to localStorage so past systems stay one click away.
What this looks like:
WHERE WE STAND(based on today’s news)
✅ AI Can Now: Hold a real-time voice chat, saying "mhmm" and pausing like a person while it listens and talks at the same time.
❌ Still Can't: Fix the hardest real software bugs as reliably as the top models. Grok 4.5 solved 53% on one bug-fixing test, while Anthropic's Fable 5 hit 70%.
✅ AI Can Now: Restyle a video you already shot, relighting it or swapping the background, from a template in a few seconds.
❌ Still Can't: Be trusted to finish a financial model without a person checking the numbers, since it can still make up facts that look right at first glance.
FROM THE WEB
RECOMMENDED LISTENING/READING/WATCHING
The Space Between Worlds by Micaiah Johnson - Book
A 2020 debut novel about a young woman from the wrong side of a walled-off Wasteland town who is one of the very few people who can traverse to alternate versions of Earth for a mysterious tech company, because you can only travel to a world where your other selves are already dead, and she has died in almost every one.
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.
