• AI Odyssey
  • Posts
  • World’s first fully autonomous AI engineer!

World’s first fully autonomous AI engineer!

Plus: Cloudflare announces firewall for AI

Hello, hello, hello! 

Welcome back to AI Odyssey’s another chapter. 

Your trusted AI insider is here to let you in on all the mind-melting AI news happening around the globe. 

Plus, some awesome AI art.

Let’s get on with it… 🤠

Meet Devin, the world’s first AI software engineer that debugs, writes, and deploys code.

Devin is a fully autonomous AI agent capable of solving complex engineering tasks by utilizing its own shell, code editor, and web browser. 

It has achieved remarkable performance on the SWE-Bench coding benchmark, correctly resolving 13.86% of real-world open-source GitHub issues unassisted, surpassing the previous state-of-the-art model's performance.

According to Cognition, Devin demonstrates the following capabilities:

  1. Learning how to use unfamiliar technologies.

  2. Contributing to mature production repositories.

  3. Training and fine-tuning its own AI models.

  4. Successfully completing real jobs on the freelancer platform - Upwork.

Devin's prowess extends beyond coding benchmarks, as it has also passed practical engineering interviews from leading AI companies, showcasing its versatility and problem-solving abilities in real-world scenarios.

The EU AI Act passed — here’s what comes next

The European Union passed the world's first major set of regulations to govern AI tech, known as the AI Act.

This landmark legislation was approved by EU lawmakers on Wednesday after years of debates.

The regulation is expected to enter into force in May after final checks and endorsement by the European Council.

Implementation will then be staggered from 2025 onwards.

How will the AI Act work?

The law's creators said it would make the tech more "human-centric."

  • The AI Act categorizes AI systems into different risk levels - unacceptable risk (banned), high risk, medium risk, and low risk. The higher the risk, the stricter the rules.

  • The AI Act bans systems designed for social scoring, emotion recognition in certain settings, and manipulation of behavior or vulnerabilities.

  • The Act pushes AI development in a direction where humans control the technology and it enables economic growth and societal progress, according to lawmakers involved.

However…

The Act does not create new laws around data collection for training AI models, leaving gray areas around the use of copyrighted data. It also exempts open-source developers and small companies.

What’s the future ahead?

Legal experts view the AI Act as a major milestone that could pave the way for other countries to follow their own AI regulations, similar to how the EU's GDPR data privacy rules set a global standard.

However, some raise concerns that the rapidly evolving AI technology could quickly make parts of the legislation outdated given the lengthy implementation timelines.

Sponsored by Kickresume

If your resume is not giving the results you need, try Kickresume. AI can CHECK your professional resume to fill in the gaps and compare it with a successful resume in order to make the necessary changes. Get your summary, work experience, and education, with a consistent tone throughout the document. No more staring at a blank screen, wondering how to improve it. 👉 Try Kickresume NOW

ChatGPT uses 17,000 times the amount of electricity than the average US household does daily

A report in The New Yorker claims that ChatGPT uses over 500,000 kilowatt-hours of electricity daily to handle 200 million requests.

That's a LOT of energy.

This is over 17,000 times more electricity than the average U.S. household uses in a day (29 kilowatt-hours).

If generative AI technology like ChatGPT was integrated into every Google search, it could consume around 29 billion kilowatt-hours per year according to estimates.

This level of electricity usage would exceed the total annual consumption of some countries like Kenya, Guatemala and Croatia.

And this is just the tip of the iceberg AI berg:

A data scientist estimated that by 2027, the entire AI industry could consume between 85 to 134 terawatt-hours annually, or around 0.5% of global electricity consumption.

It makes you wonder what the footprint will be as this technology keeps growing… On the bright side, this news might inspire some creative solutions for more sustainable AI development!

AI art to sip your coffee with…

Prompt: 

**fleeting pergamot scintillating shiny plastacine melo ties in surreal looking glass alchemy

Well, that looks like a glass of heaven 🍷

💣Latest AI hits you can’t miss:

  • Anti-AI sentiment gets big applause at SXSW 2024 as moviemaker dubs AI cheerleading as ‘terrifying bullsh**’. (link)

  • Amazon will let sellers paste a link so AI can make a product page. (link)

  • Microsoft will make its generative artificial intelligence (GenAI)-powered security tool, Copilot for Security, generally available worldwide beginning April 1. (link)

  • AI image-generator Midjourney blocks images of Biden and Trump as the election looms. (link)

  • Anthropic just released a Claude 3 AI prompt library. (link)

  • OpenAI CTO Mira Murati tells The Wall Street Journal that Sora will eventually incorporate sound as well. (link)

  • DoorDash’s new AI-powered ‘SafeChat+’ tool automatically detects verbal abuse. (link)

  • Google won’t let you use its Gemini AI to answer questions about an upcoming election in your country. (link)

  • Perplexity brings Yelp data to its chatbot. (link)

  • Cloudflare announces Firewall for AI, aims to provide much-needed security controls for LLMs. (link)

🧠 Quote of the week:

AI may be “smarter than all humans combined by 2029”, says Musk

Elon Musk, CEO of SpaceX and Tesla and owner of X, formerly known as Twitter, attends the Viva Technology conference 

That’s a wrap! See you next week on Monday for more AI madness. 😜