AI Top-of-Mind for 7.3.24 — CriticGPT
Today: CriticGPT, AI warfare in Ukraine, Meta Chameleon, and what powers the Las Vegas Sphere? Note that I’ll be taking the 4th off in honor of US Independence Day.
Top-of-Mind is OpenAI’s new ‘CriticGPT,’ with ‘Marktechpost’ offering details:
· CriticGPT’s primary purpose is to produce thorough criticisms that draw attention to mistakes, especially in code outputs. This model has been created to overcome the inherent limitations of human review in RLHF. It offers a scalable supervision mechanism that improves the precision and dependability of AI systems.
· CriticGPT has proven to be remarkably effective in enhancing the assessment procedure. In experiments, human reviewers who examined ChatGPT’s code outputs with CriticGPT performed 60% better than those who did not receive such assistance.
Link to the original OpenAI paper.
And the latest on the AI weaponry front, and it isn’t only robot dogs. The ‘NY Times’ reports on how the Ukraine war has changed the face of ground warfare. From the article:
· Vyriy is just one of many Ukrainian companies working on a major leap forward in the weaponization of consumer technology, driven by the war with Russia. The pressure to outthink the enemy, along with huge flows of investment, donations and government contracts, has turned Ukraine into a Silicon Valley for autonomous drones and other weaponry.
· What the companies are creating is technology that makes human judgment about targeting and firing increasingly tangential. The widespread availability of off-the-shelf devices, easy-to-design software, powerful automation algorithms and specialized artificial intelligence microchips has pushed a deadly innovation race into uncharted territory, fueling a potential new era of killer robots.
And the key issue:
Major questions loom about what level of automation is acceptable. For now, the drones require a pilot to lock onto a target, keeping a “human in the loop” — a phrase often invoked by policymakers and A.I. ethicists.
On the model front, ‘Toms Guide’ describes Meta Chameleon, an open-source GPT-4o class LLM that handles both text and images. From the article:
In other words, you could take a picture of the contents of your fridge and ask it what you can cook using only the ingredients you have. This is something not possible with the Llama generation of AI models and brings open source closer to the higher profile mainstream vision models from OpenAI and Google.
A link to the original Meta paper.
Next up, if you are familiar with the Las Vegas sphere, what technology powers it? The ‘Mercury News’ dives into a lesser-known company ‘Disguise’ that powers all the visuals. With more powerful computing, assisted by AI, we’ll expect to see more venues like this. An important point is that Disguise’s technology processes and displays the visuals, with companies like ILM creating the content. From the lead-in to the article:
· When the rock band Phish began their 2004 track A Song I Heard the Ocean Sing at the Sphere in April, more than 18,000 fans stared, mouths agape, as the quartet from Vermont seemingly jammed in the middle of a coral reef.
· Fish — not to be confused with Phish — swam from one side of the stage to the other amid giant, pulsating jellyfish as the whole dome transformed into a trippy underwater experience. Human bodies bobbed up and down in the water amid radiant flashes of red and green. Tall plants shot up from the sea floor. The scene was just one of dozens of visuals displayed inside and on the exterior of the glowing $2.3 billion dome in Las Vegas — the most talked-about concert venue in the world.
Lastly, and not exactly AI, but something that nags on all of us…. the inefficiencies of using different message apps and why they are not compatible. Larry Magid writing for the ‘Mercury News’ dives into the difficulty of navigating between iMessage, Google Chat, WhatsApp, and the various social platforms. And it isn’t just the ‘blue-green’ thing that impacts the user experience between Apple and Android. I like his summary:
In some ways, today’s messaging environment reminds me of what it was like before the internet when you had to be on the same email system as the person you were messaging. Services like MCI Mail, AOL, CompuServe and Prodigy weren’t interconnected. Internet email changed all that but now we’re back to a gaggle of unconnected messaging apps.
