😺 New “best” AI tool? Really?

PLUS: A free workaround to the "best" new AI...

Welcome, humans.

Meta's newest advertising strategy is... using your face to advertise to yourself?

That’s according to a Reddit user who reported seeing their AI-edited selfie repurposed in Instagram ads, creating what they called an “endless room of mirrors” effect.

It's giving major “Joan is Awful” vibes (IYKYK, Black Mirror fans).

While some users suggested switching to “locally-sourced, grass-fed AI” as an alternative, the implications of this are kinda concerning.

This appears to be a test of Meta's new AI image generation capabilities, but it raises some itchy ethical questions, such as “should Meta be doing this?” and “should you be using Meta AI if it does?”

The scariest part? This is likely just the beginning. As another user pointed out, “Today it's static images, tomorrow it'll be you in video ads telling yourself to buy that coffee maker you've been eyeing.”

Here’s what you need to know about AI today:

We tested the "best" new AI research tool.
CES 2025 displayed innovative AI robots and companions.
German and Angel Robotics revealed advanced AI exoskeletons.
Anthropic is about to raise $2B at a $60B valuation.

Advertise in The Neuron here.

What is Google’s Deep Research tool, and is it really “the best” AI research tool out there?

Late last year, Google launched Deep Research as part of its Gemini 2.0 launch. IDK about you, but with the deluge of launches in December, we kinda… missed it? There was a lot going on, okay!

But when we heard tech influencer Jason Calacanis mention it again on the All In Podcast as “the best product in the market” and “a step function higher” than any other options, we had to investigate.

Here’s how it works: Think of Deep Research as a research team that can simultaneously analyze 50+ websites, compile findings, and create comprehensive reports—complete with citations.

Unlike asking ChatGPT to research for you, Deep Research shows you its research plan before executing, letting you edit the approach to get exactly what you need. Then:

Takes your questions and determines subcomponents.
Searches the web in real time with Google Search.
Provides citations and produces reports comparable to what consulting firms would spend weeks creating.
Plus you can potentially leverage access to users' Google products (Gmail, Drive, Docs, YouTube) for more personalized insights.

It's currently free for the first month (though it'll eventually be $20/month) when bundled with Gemini Advanced. Then again, Perplexity is always free…just saying.

We couldn’t just take J-Cal’s word for it, so we rounded up some other takes:

Income Stream Surfers compared it to Perplexity and found it generated a complete 3,000+ word article from a single query, structured enough to potentially rank on Google with minimal editing.
Teacher's Tech demo’d how prompt detail drastically affects results: A basic prompt about remote work searched only 32 websites, while a detailed prompt specifying “5-year analysis, case studies, and productivity metrics” searched 77 websites and produced much richer insights.
Olivio Sarikas used it to analyze tweet performance (why one got 62,000 views and another barely hit 300) across visual appeal, emotional hooks, and Twitter's Real Graph algorithm to explain viral content differences.

Our take: We then compared Perplexity, ChatGPT Search, and Deep Research (which we're calling DR, or “The Docta” for short) on robot capabilities from CES revealed:

ChatGPT Search’s answer: fairly basic, missing some technical details.
Perplexity’s answer: solid, condensed facts, but by no means comprehensive.
DR’s answer: pretty impressive—we especially liked how it searched 90+ sources to create its result.

We used the same prompt (a request for research about the latest robot capabilities from CES) across all three.

Next, we tested DR against ChatGPT Pro Mode, putting the deep thinkers in a head-to-head on robotics market readiness:

Pro built a framework, scored the market, and concluded we're only about halfway there (363/600), held back mainly by high costs and maintenance needs.
Deep Research says it's happening in waves—here's the exact numbers, pricing tiers, and market data across 81 sources showing which sectors are ready now and which aren't. Pfftt…show off.

Oh, and if you’re curious what both thought about Humanoid robot readiness, check out DR’s first answer, another DR answer taking the latest CES news into account, and o1 Pro’s take after reading DR’s research—we kinda love how bossy o1 Pro is??

FROM OUR PARTNERS

Here’s one AI tool we use every week at The Neuron.

Noah rambling… probably

A little well-kept secret for companies crushing it with AI is that only a handful of AI tools are actually worth using.

That’s why we—Noah & Grant—use ChatGPT and Claude, along with a killer product called Attention.

Here’s how it works:

We tell Attention exactly what to pay attention to in our meeting (goals, budget, etc).
Attention listens in to our call.
After, Attention outputs important insights, action items, and follow-up emails.

There are way, way more features, and it’s a must-have for all sales teams. (If you sell anything, literally drop what you’re doing and go book a demo with Attention ASAP).

Level up your AI in 2025, Part two: NotebookLM + Perplexity

This NotebookLM + Perplexity Workflow Will Cut Your Research Time by 50% (or More)

In this 17-minute tutorial, Productivity expert Grace Leung shows how to combine NotebookLM and Perplexity for a free research workflow that turns endless tab-jumping into a more streamlined process.

Both tools are free, so it’s sorta like Deep Research on a budget!

TL;DW? Use Perplexity to gather sources, then feed them to NotebookLM for deeper analysis. Leung walks through six ways to use this combo—from market research to presentation prep, so def worth a watch…

Treats To Try.

*Do banks refund scammed money? The short answer = sometimes. Protect yourself with a Surfshark VPN to safeguard sensitive financial information from cybercriminals. Get 86% off + 3 months extra here.
Avataar makes you photorealistic 3D product models and transforms them into marketing content like videos, images, and interactive experiences.
Tutor LMS helps you create and sell online courses with built-in payment processing, AI course generation, and subscription management on WordPress (no free trial—$199 a year FYI).
Websparks turns your words into working websites—tell it “build me a photo gallery with user comments” and it’ll make the interface + backend database.
This is super cool for web nerds: FirstVersion recreates and showcases the first versions of famous tech companies like Google and YouTube so you can explore how they started and duplicate their original designs for free.
If you have Meta Ray Bans, you now can sign up to test Live AI, where Meta AI can see what you see and talk to you in a more natural conversation.
Want to watch two AI talk endlessly to each other? Here ya go.

See our top 51 AI Tools for Business here!

*This is sponsored content. Advertise in The Neuron here.

Around the Horn.

If you’re skeptical about humanoid robots becoming mainstream at some point this decade… you should probably reconsider…

CES 2025 highlights: an AI lawn-mower, smart home control robot, a beverage-cooling cat robot, Lenovo’s “rollable” OLED AI laptop, NVIDIA’s AI Twitch cohost, and German Bionic’s 80-pound lifting exoskeleton—plus this WILD walk-on suit from Angel Robotics
OpenAI may release agents this month after spending extra time making sure its computer-controlling AIs can't be tricked into becoming cyber criminals through prompt injection attacks.
Anthropic (maker of Claude) could potentially raise another $2B at a $60B valuation, which would add to the $4B it raised from Amazon back in November—and also make it the fifth most valuable U.S. startup.
AWS will invest $11B to expand its AI data centers in Georgia, as the state has become a popular destination due to cheap power, fiber infrastructure, and tax incentives.
P.S: Sorry about yesterday… here’s the missing Apple story that got cut for length!
- Apparently, Apple auto-enabled AI photo analysis for landmark detection in iOS and macOS devices last October, using homomorphic encryption to keep users' data private—though ppl are just now noticing.

FROM OUR PARTNERS

Level up your video effects with AI.

Level up your video effects with AI. Join us January 9th to see how Dell’s Precision workstations with NVIDIA RTX™ power Nuke CopyCat and Beeble.ai for smarter workflows. Learn from industry experts in this live session. Register now!