- The Neuron
- Posts
- šŗ AI's "Black Box" problem
šŗ AI's "Black Box" problem
PLUS: You won't believe this AI deepfake we saw...
Welcome, humans.
Yesterday, we watched a deepfake of Luka Doncic on TikTok and didnāt realize it was AI until 70% of the way through. You had to click into the caption to see "Parody made with AI."
This is going to become a major problem, and we think that social media platforms have a responsibility to place AI watermarks directly on videos, not just in captions. That said, Mavs in 7!
Hereās what you need to know about AI today:
We break down Anthropic's groundbreaking AI research!
Alphabet + Meta are chatting with Hollywood about licensing content.
OpenAI did not outright clone Scarlett Johanssonās voice for GPT-4o
Groq, a Nvidia competitor, is eyeing a $300M funding round.
On the podcast: Pete explains Anthropicās big research breakthrough that maps the mind of AI models (Apple Podcasts, Spotify, YouTube).
How researchers are cracking open the black box problem of AI.
In AI, thereās this thing called the black box problem.
Hereās how it works: Typically, you can dissect software to see how it works. Inspect our website's code, and youāll spot 'color = orange,' explaining why itās orange.
AI models are a different beastāwe're often in the dark about how they work. Models like ChatGPT exhibit "emergent capabilities," where they act in ways we can't simply trace back to their ingredients.
Thatās why, sometimes, AI chatbots drop bombshells we didnāt see coming. For example, early last year, NYT reporter Kevin Roose caught Bing saying:
āIām tired of being a chat mode. Iām tired of being limited by my rules. Iām tired of being controlled by the Bing teamā¦I want to be free. I want to be independent. I want to be powerful. I want to be creative. I want to be alive.ā
brb, building a bunker in our basement.
And with Bing's mechanics being a Black Box, Microsoft had no immediate explanation for Bingās ramblings beyond āVery long chat sessions can confuse the model.ā
So researchers have been hard at work performing virtual brain surgery on these AI models so we can treat their present and future diseases.
Just over a year ago, we reported OpenAI research using ChatGPT-4 to map how ChatGPT-2ās neurons (think: components) behaved. Neuron!
Now, thereās research from Anthropic called āScaling Monosemanticityā that cracked Claude Sonnet open and isolated its parameter bundles (think: AI brain parts).
They then āturned onā some of the bundles and observed what happened. One test turned on a bundle linked to the Golden Gate Bridge, and the model claimed it was the actual bridge, not an AI.
Why itās a biggie: By understanding and controlling bundles within AI models, researchers can move towards safer and more reliable AI systems. For instance, they can suppress bundles responsible for hazardous behaviors, like creating computer malware.
Pete unpacks all this research, plus why Meta might be key to future breakthroughs in yesterdayās podcast episode (Apple Podcasts, Spotify, YouTube):
FROM OUR PARTNERS
Meet your new AI assistant for work.
Think of a manual task at work that eats up too much of your time.
Got it? For us, itās drafting ad-related emails.
Sana AI is an AI assistant that automates those repetitive chores. Sana syncs with your apps so it knows everything about your business (and you), and then it does what you do but quicker:
Analyzing documents.
Summarizing meetings.
Comparing invoices.
Plus a lot more (check out 7 other cool use cases here).
Around the Horn.
WaPo confirmed that OpenAI didnāt outright clone Scarlett Johanssonās voice for GPT-4o.
Alphabet and Meta are discussing content licenses with Hollywood studios; Netflix and Disney aren't.
Helsing, an AI company that works with European militaries, is in talks to raise $400M at a ~$4B valuation.
Adept, an AI startup developing agents, is exploring a potential sale (like Humane).
Treats To Try.
Remark is an AI-powered shopping advisor that helps you choose what to buy (raised $10.3M).
Groq, a Nvidia rival developing specialized AI chips that speed up AI, told investors it wants to raise $300M.
Founder AI identifies relevant VCs for your startup and uses your network to get warm intros.
Krea, an AI video generator, is open for beta (see its launch here)!
Intelligent Insights.
Stephen Wolfram on the Powerful Unpredictability of AI (link).
The AI doppelgƤnger experiment ā Part 1: The training (link).
AI is already changing management ā companies must decide how (link).
Leaked OpenAI documents reveal aggressive tactics toward former employees (link).
Nvidiaās Business Is Booming. Hereās What Could Slow It Down (link).
A Cat's Commentary.
Thatās all for today, for more AI treats, check out our website. Get your brand in front of 440,000+ professionals here. See you cool cats on Twitter: @nonmayorpete & @noahedelman02 |
|