- Bot Eat Brain
- Posts
- 🧠 Here's why the Space Force banned ChatGPT
🧠 Here's why the Space Force banned ChatGPT
PLUS: Making friends is hard. Buy AI instead.
Michael Parrish
October 13, 2023
TOGETHER WITH
Good morning, human brains. Welcome back to your daily munch of AI news.
Here’s what’s on the menu today:
Why did the Space Force ban ChatGPT? 🧐 🚀
Allegedly, it’s to protect agency data while AI policies are created.
Adobe claims you can’t edit out its watermark 🍯 🌽
It unveiled its “icon of transparency” to label AI content.
No friends? None needed, buy AI instead 💔 💻
Character.AI released an AI group chat functionality on Wednesday.
MAIN COURSE
Here’s why the Space Force banned ChatGPT 🧐 🚀
In August, we reported on The Department of Defense’s “Generative AI Task Force.”
Those clowns did what, now?
An official memo was sent to the Guardian Workforce (Space Force members) that stated they “are not authorized” to use web-based generative AI tools on any government system unless specifically approved.
But… Why..?
Lisa Costa, the Space Force’s Chief Technology and Innovation Officer, said it’s to protect agency and personnel data while officials set AI policies.
So, the Space Force is using AI unsafely?
Costa claims because of cybersecurity, data handling, and procurement uncertainties, “responsible” adoption is required.
Who does this affect?
At least 500 people who used an AI platform, Ask Sage.
So, Ask Sage is unsafe?
Nicolas Chaillan, the former chief software officer for the Space Force and Air Force, founded Ask Sage. It aims to provide a secure AI platform that works with many different LLMs.
He claimed Ask Sage was already approved by the Defense Department and the Air Force. He also said the ban would “put us years behind China.”
FROM OUR PARTNERS
Apply to hundreds of jobs in 30 seconds. 🤝
In the age of AI, no one should spend hours clicking "upload resume" over and over again.
You should be spending your time prioritizing companies that have already shown interest in you, not waving your hand just trying to be seen.
ApplyAll maximizes your chance of getting an interview by using AI to apply to hundreds of relevant jobs tailored just for you.
Their bulk-apply AI applies you to 200 relevant jobs to supercharge your job search, whether you're looking to work at a startup or a big tech company.
Game the application process
Accelerate your job search without pouring time and effort into the application process.
Get interviews in your field
Bulk applying leads to more interviews, and they claim a 99% success rate to prove it.
All you need is a resume
AI takes care of the rest. ApplyAll saves you hours, even days, by automating the repetitive stuff.
Accelerate your job search, save yourself hundreds of hours, and get your dream job.
BUZZWORD OF THE DAY
AI safety
The focus of preventing accidents, misuse, or other harmful consequences that could result from AI systems.
This could be anything to making sure LLMs are accurate, to safeguarding humanity from uncontrollable AI scenarios.
Examples:
ChatGPT’s web browsing was disabled due to potential hacks and to avoid spreading false or harmful information.
In July, The White House brokered a voluntary AI safety agreement with the top firms.
SIDE SALAD
Adobe marks its territory 🍯 🌽
Adobe gets territorial, and not in the hot way.
Yesterday, we covered Adobe’s new AI image tool — Project Fast Fill. It’s like Photoshop’s generative fill feature, but for videos.
It unveiled its “icon of transparency” — its new symbol to label AI-generated content.
Icon of what, now?
Adobe compares the symbol to a nutrition label, but it’s not a very accurate metaphor. Basically, it just indicates the origin of the media.
How does it work?
When you hover over a photo with the icon enabled, it reveals details like content ownership, the AI tool used, and more.
I can just edit it out, right?
Adobe claims it’s embedded into the metadata, so it can’t be easily removed or edited. You can probably find a workaround, though.
Here’s what it looks like:
So, all AI content is watermarked now?
On Tuesday, we covered Meta’s AI watermarking tool, Stable Signature. Though it’s an improvement of DeepMind’s SynthID, humans will always find a way around these tools.
A LITTLE SOMETHING EXTRA
Don’t make real friends, just use AI 💔 💻
“All the lonely people…”
Character.AI is an AI chatbot startup backed by a16z and founded by ex-Google AI researchers.
It just introduced a “Group Chat” experience.
I don’t have friends, what is it?
It lets you have AI companions with unique personalities and tools to design your own personalities.
Elon, where did my life go wrong?
What if I miraculously make a friend?
You can form group chats with your friends and add AI companions.
Friends for free? Sounds too good to be true.
It is. It costs $9.99 per month.
MEMES FOR DESSERT
YOUR DAILY MUNCH
Think Pieces
Ethan Mollick answers FAQs about AI. The Wharton University professor talks about AI’s safety concerns, detection challenges, and more.
How advanced are GPT-4’s reasoning capabilities? What are its blind spots, and does it have the capacity to become sentient?
How do you chat with your data on LangChain? This Coursera project will teach you all you need to get started in an hour.
Startup News
Anysphere gets $8 million from OpenAI, Nat Friedman, and more. The money will go to building an “AI-first” coding editor.
Conveyor raises $12.5 million in funding. It’s to further develop a security portal and an LLM that understands the structure of security questionnaires.
TitanML, a London startup, raises $2.8 million. The money will aid the development and deployment of its LLM, “Titan Takeoff.”
Research
LLMs for Dummies — contains 4 frameworks for LLM applications and an easy-to-read guide to get started.
Mistral 7B — the official paper of the AI model that beats many models twice its size.
EVEscape — an approach for the continuous evaluation of the risk posed by emerging variants that could lead to future pandemics.
Tools
Moonhub AI — an AI-powered hiring tool used by billion-dollar companies.
Blaze — automate and streamline your content creation process.
Dashtoon — an AI-powered comic creation tool.
Relay — an AI-powered workflow automation tool.
RECOMMENDED READING
💡 Web3Daily — Web3 and crypto news, translated into plain English. A hype-free, 5-minute daily read, designed to be your gateway drug to Web3.
Check out their latest article here:
TWEET OF THE DAY
What if they had sneakers in 336 B.C.? Renowned Wharton University professor tweets some images generated by DALL-E 3 powered by GPT-4. More on DALL-E 3, here.
Tag us on Twitter @BotEatBrain for a chance to be featured here tomorrow.
AI ART-SHOW
Until next time 🤖😋🧠
What'd you think of today's newsletter? |