- Ben's Bites
- Posts
- Rogue AI - concern or cope
Rogue AI - concern or cope
PLUS: give to get access, skills in AI world, Bing chat ads
Hey folks, strap in. The debate between the dangers and benefits of AI is picking up heat, EEK!
Let’s get to it.
p.s. I’ve made ‘Ben’s List’ - my top recommendations of tools, investors, papers, blog posts, and more. 100 links total right now, and I’ll add to it each month. It’ll be a paid access thing released here in the next ~2 weeks - I’m taking pre-orders now, for a discount.
p.p.s. I’d love it if you could fill out this quick survey to learn a bit more about who’s reading, how I can improve etc. I’ll pick a few entries to get free access to the list above and throw in a $100 gift card.
From our sponsor:
Use AI directly in your notes.
Reflect is a note-taking app that mimics the way your brain works. It has integrated GPT as an AI assistant.
Edit and improve your writing
Organize your notes into to-do lists
Summarize key takeaways
and much more. Bonus: Use OpenAI's Whisper to transcribe voice notes.
You can sign up for a Reflect and test the AI assistant here.
🤌 Our Picks
Highlights if you've only got 2 minutes
1/
Remember the pause AI letter we mentioned yesterday? Well, it has certainly kicked up a reeeaaal big fuss on the AI scene, and we’re here to keep you updated.
For those who don’t already know, Future of Life Institute created a signable petition to halt the training of large language models for a whole six months. And Gary Marcus (Neural Science prof. at NYU, and often a critic of the pace of AI advancements) heavily publicised it.
Now Andrew Ng, the tech world’s equivalent of David Attenborough and Coursera’s Co-founder, has waded in to put his foot down. He argues that having governments pause emerging technology they don’t understand is anti-competitive at best, and sets a terrible precedent at worst. It’s simply not feasible, and even if it was, it won’t help us get anywhere - regulations around transparency and auditing would help us navigate explosive tech far more effectively.
2/
The doom and gloom trend doesn’t end there… Time magazine posted an opinion piece saying the letter hasn’t gone far enough! Read it without the paywall here.
Eliezer Yodkowsky starts the piece with, “I refrained from signing because I think the letter is understating the seriousness of the situation and asking for too little to solve it.” Acc. to him, AI will kill us all, so he is arguing that we simply are not prepared and it’s time to just “shut it all down.” Sheesh!
It’s worth mentioning some balanced takeaways as well. This piece from Jon Stokes on AI Safety - A technical & ethnographic overview is worth a read.
This topic is starting to gain enough attention from big enough names that we might just see the higher powers step in soon… luckily for us; those aren’t AI machines just yet 😉.
3/
Now, I’m glad that I’m not quite old enough to remember Jigsaw, a start-up founded in 2004 (I was 14), and David Sacks argues that Jigsaw’s business model is overdue for a revival with this current wave of AI. The SaaS industry lives by the frameworks from David, so worth reading this.
The ‘Give-to-Get’ model, where people could upload their business contact and receive access to contacts in response. See where this is going…? You got it. Why not apply the Give-to-Get model to datasets that can be used to train AIs? Hand over some data, and get access to powerful models in return. With baked-in earning points for activity that allows ‘purchase’ access to more features. As the lifecycle of the framework goes forward, the power of the models grows and it amplifies itself. I love it!
4/
It’s time to say hello to another awesome-looking generative art model, and this one is from Playground AI. Creatively called Playground v1, they claim that it’s able to improve the overall HDR quality of the images generated, and I have to say they do look damn crisp and detailed. You can try it out for free on their website, where you’ll see that you can even choose a photo style to dominate the image - and they come with snazzy descriptions like “radiant symmetry”, “ominous escape”, and “neon mecha” (whatever that is!).
🛠️ Cool Tools
Product launches, updates and demos
Story AI - Write your own story with artificial intelligence. (link)
GPT Calculator - Easily calculate the token count and cost of your GPT Prompt. (link)
Libraria - Autonomous platform to embed custom assistants from your own data. (link)
The AI Sorting Hat - Sort your terrible drawings into their Hogwarts house with AI. (link)
Anime from Revel - Free consumer service that enables animated avatar creation from a single pic. (link)
Question answer assistant for all 121 episodes of the All In Pod. (link)
Taskade - Generate task lists, mind maps, and structured notes, right inside the app. (link)
Vocode - An open-source library for building voice-based LLM applications. (link)
60-sec site - Create a custom landing page for your app in 60 seconds. (link)
Talk to ChatGPT over the phone! (link)
Kapa AI API - Answer technical questions for developers automatically. (link)
Hugging Face Unity API, an open-source plugin that streamlines the use of HF inference API in Unity projects. (link)
Demo from Respell - A single lesson being instantly converted into a version for visual learners, logical learners, etc. (link)
Instant Chapters - Timestamped chapters for your youtube videos with only one click. (link)
Uncle rabbit - The first conversational holographic AI. Warning - it’s too cute. (link)
ChatGPT wrapper to keep using ChatGPT even when there’s an outage. (link)
Asksumo - Chat with AskSumo to find cost-effective software alternatives. (link)
Pico - Use GPT4 to instantly build simple, shareable web apps. (link)
pssssst…I’d love it if you could fill out this quick survey to learn a bit more about who’s reading, how I can improve etc.
🤓 Miscellaneous
News, podcasts, videos, blogs etc
AI could replace the equivalent of 300 million jobs - report from Goldman Sachs. (link)
Replit and Google, the Google perspective, Twitter’s bad idea. (link)
What we still don’t know about is how AI is trained. (link)
In the age of AI, don't let your skills atrophy. (link)
Startups vs incumbents in AI. (link)
ChatGPT is a knowledgeable but inexperienced solver. (link)
Bing Chat ads are getting to more users. PS: They were there from day one. (link)
Top 10 projects from the AGI House hackathon with 170 founders and researchers. (link)
Video and transcript of fireside chat with Clem Delangue, CEO of Hugging Face. (link)
10 UK companies at the forefront of AI - from DSIT UK’s white paper on AI. (link)
Alphabet’s Google and DeepMind pause grudges, join forces to chase OpenAI. (link)
Rise of open source AI models trying to match ChatGPT. (link)
Forget Taylor Swift, OpenAI is doing a tour. Register here.
🎓 Learn
How-to’s and resources
👋 Too many links?! I created a database for all links mentioned in these emails. Refer 1 friend using this link, and I'll send over the link database.
📰 Unclassifieds
Short, sponsored links
Sam Parr and I are both, independently, looking for a solution to matchmaking using AI. Know anyone?
DeepMedia is organising a remote Hackathon this Friday, 31st March. Details here.
Inner friend contest from Novo - Prompt engineer a bot to become your best friend and win $1000, deadline Apr 6th. (link)
JAX/Diffusers community sprint - Work on fun and creative diffusion models using JAX and Diffusers. (link)
Add your open-source AI project or research for a chance to be showcased at the Hugging Face event in San Francisco today. (link)
Have a product, service, job, event, newsletter, app, book, movie, tool, or anything you'd like to share with over 60k subscribers?
Advertise
🖼 AI Images of the day
Funny memes and pics from around the web
Send this to 1 AI-curious friend and receive my AI project tracker database! Use this link.
How was today's email? |
Reply