- Lore
- Posts
- Lore Issue #39: OpenAI VS Claude Heats Up
Lore Issue #39: OpenAI VS Claude Heats Up
After a month of things being relatively slow in AI, there are signs that things are heating up again.
This week OpenAI introduced Code Interpreter for all ChatGPT Plus users. As we mentioned in Lore Issue #30, Code Interpreter is one of the most exciting developments in AI that could significantly enhance the work of data scientists.
Meanwhile, Anthropic released Claude 2.0, a major upgrade to the ChatGPT competitor. They’re quickly closing the gap with GPT-4, and arguably they have a better free offering now than GPT3.5.
I’m excited to see the increasing competitive pressure on OpenAI, which I expect will result in GPT4.5, with multi-modal support coming sooner.
OpenAI Releases Code Interpreter
Now all ChatGPT Plus users have access to Code Interpreter!
With Code Interpreter, you can analyze data, create charts & graphs and even run Python code and install Python packages.
To turn it on, select Settings → Beta features and ensure Code Interpreter is on.
If you’d like to see examples, we previously created a Twitter thread. Also, our friend MattVidPro created a great video showcasing why Code Interpreter is a big deal.
Anthropic Debuts Claude 2, Their Next-Gen AI Model
AI startup Anthropic has unveiled its latest large-language model, Claude 2, with considerable improvements in coding, math, and reasoning capabilities over its predecessor.
The enhancements in Claude 2's performance are noteworthy. For instance, it scored 71.2% on a Python coding test, a remarkable improvement from the 56% its predecessor achieved. Additionally, its ability to comprehend and process language doubled, enabling it to handle the equivalent volume of an epic novel.
With this release, OpenAI now has a serious competitor. If you compare free offerings from OpenAI with GPT3.5 and Claude 2, Claude 2 is superior in almost every way. Most notably, it has a 100k context window. Meaning you can upload 100s of pages.
The consumer version of Claude 2 is currently free and generally available in the US and UK.
US Military Leverages Generative AI for Enhanced Planning and Decision-making
Breaking from its traditionally slow tech adoption, the US military is exploring generative AI for strategic planning and decision-making. As Bloomberg reports, they're trialing five Language Learning Models (LLMs), trained on classified data.
This strategy goes beyond data processing, aiming to incorporate AI into aspects of military planning, sensor analysis, and firepower decisions.
Information on the tested LLMs is limited, but trials include ScaleAI's defense-focused Donovan platform. Microsoft and OpenAI, due to their established Defense Department relationships, may also be involved.
Other news:
Pika Labs
This week, the coolest things we found were from an unlikely source - Pika Labs, a new AI Video startup.
Despite being significantly less funded than its competitor, Runway, Pika Labs already has a promising start. In terms of video quality, it seems to be comparable to or even better than Gen-2 by Runway, which many people found to be a bit of a disappointment after all of the hype.
Check out some of the videos people are making with Pika Labs. They’re pretty incredible.
Playing with @pika_labs image to video. First is the video second the midjourney image I used to prompt it. Video completed with edited on AI synched voice and sound fx. twitter.com/i/web/status/1…
— Uncanny Harry AI (@Uncanny_Harry)
1:45 AM • Jul 12, 2023
Just got a camera movement from an input image with @pika_lab new image to video.
Compare how close the start frame is to the image below.👇
Exciting!!
#aivideo#AIcommunity#midjourneyV52— ChristianF (@ChristianF369)
1:16 AM • Jul 12, 2023
Pika labs releases image-conditioned video generation, upload an image, and the model will animate the image, prompt is “a girl in the wind”
twitter.com/i/web/status/1…
— AK (@_akhaliq)
12:47 AM • Jul 12, 2023
Quivr: Dump all your files and chat with it using your Generative AI Second Brain using LLMs. (Link)
MiniStudio: App that allows children to bring their own virtual characters to life. (Link)
PhantomBuster: Manage your leads, get data insights, and write personalized messages with AI. (Link)
AgentOps: Tools to make agents that work. Build your next agent with graphs, monitoring, and replay analytics. Join the waitlist. (Link)
Find more tools at FutureTools.io.
That's it!
See you next week!
If you want to support Lore, please:
Share this newsletter
Follow us on Twitter @NathanLands & LoreAiNews.
Join our Generative AI community on LinkedIn.