Anthropic has launched the latest version of its AI chatbot named Claude 2 and is only currently offering the service for free in the US and UK. Here’s how the chatbot compares with OpenAI’s ChatGPT.
Anthropic is led by Dario Amodei – the former OpenAI VP of research who reportedly quit the company over issues related to commercialization. In an interview, he told Forbes that Claude 2 is a “bigger jump” than a “gargantuan leap” over the previous version Claude 1.3, and termed it as part of the “evolutionary process.”
He added, “You have to delineate things somewhere, and we felt that this was a substantial enough improvement, even if it’s part of a general progression.”
Introducing Claude 2! Our latest model has improved performance in coding, math and reasoning. It can produce longer responses, and is available in a new public-facing beta website at https://t.co/uLbS2JNczH in the US and UK. pic.twitter.com/jSkvbXnqLd
— Anthropic (@AnthropicAI) July 11, 2023
To be sure, Claude 2 has shown meaningful improvement as compared to the previous version and scored 76.5% on the Bar exam’s multiple-choice section as compared to 73% in the previous version.
In the Python coding test, Claude scored 71% – up 15 percentage points from the earlier version. Similarly, the model improved its middle school math quiz grade to 88% from 85.2%.
While the previous version could analyze a prompt of only about 75,000 words, the new model can handle twice that.
OpenAI also raised funds around the same time at a reported valuation of between $27 billion-$29 billion.
AI Funding Has Been Tepid in 2023 Despite the Boom
Startup funding activity has been quite tepid this year. While there is a global euphoria over AI and most tech companies are either investing in enhancing their AI capabilities or acquiring AI startups, the venture capital (VC) funding data tells a different story.
Global AI venture funding plunged in Q1 2023 in line with the slump in the overall VC industry. According to Crunchbase, global VC funding into AI startups fell 43% YoY to $5.4 billion in the first quarter of the year.
However, in what looks encouraging for AI companies, Crunchbase noted that eight of the 38 unicorns in the first five months of this year are AI companies.
Meanwhile, after the launch of Claude 2, many are wondering how the model compares with ChatGPT.
According to Decrypt, while ChatGPT Plus supports 8,192 tokens, Claude 2 supports 100,000 tokens of context. While an OpenAI version supports 32,000 tokens it is not used by ChatGPT.
How Claude 2 Compares with ChatGPT
In terms of pricing, while ChatGPT offers the versions upto GPT-3.5 for free, it charges $20 per month for the GPT-4 version (ChatGPT Plus). For now, Claude 2 is free but the version is only available in the US and the UK.
In contrast, ChatGPT is available in most jurisdictions barring the usual suspects like China and North Korea.
Anthropic literally killed every "chatPDF" wrapper with Claude2 – you can upload files now
I tried Tesla's latest q1 update, and asked
"What are the key takeaways from this q1 update? Can you make any forecasts for the price?"
It was able to answer it flawlessly, with sources pic.twitter.com/oaA3Bgj8GB
— Sully (@SullyOmarr) July 11, 2023
Also, unlike ChatGPT, Claude supports browsing through VPN. In general, Claude is less prone to harmful responses than ChatGPT and Anthropic claims that Claude 2 is twice as good as its previous version when it comes to limiting harmful outputs.
Amodei said, “It’s almost a certainty that someone’s going to find some new jailbreak in the model, someone’s going to find some bad hallucination. There are too many things you can say to the model, and too many things the model can say back.” He added, “These things are never going to be perfect.”
AI Execs Have Warned of Existential Threat
Notably, several AI executives have warned that the technology could be an existential threat – even as their reasons might not be fully altruistic.
Musk is among the tech leaders who have been constantly warning about AI risks. Speaking at the Wall Street Journal CEO Summit last month, the Tesla CEO said, “It’s a small likelihood of annihilating humanity, but it’s not zero,” while adding there is a “non-zero chance of [AI] going full ‘Terminator’.”
To be sure, Musk has warned of AI risks before also and termed it “dangerous.” He has been lobbying aggressively for AI regulations and in April met Senate Majority Leader Chuck Schumer and other US lawmakers to discuss AI regulations.
After his meeting, the White House courted CEOs of OpenAI, Google, Microsoft, and Anthropic – where they discussed AI with Vice President Kamala Harris.
Meanwhile, these are still early days for Claude 2 and so far, it seems the model is better than ChatGPT when analyzing something which entails long texts – given its mammoth token limit.
On a holistic level though ChatGPT is still probably better. However, both Claude and ChatGPT might get better with future updates as AI continues to make new advancements.
Related Stock News and Analysis
- Trends in Tech Layoffs Suggest the Bottom is Near: More Tech Companies Are Cutting Jobs but Fewer Tech Workers Are Being Laid Off
What's the Best Crypto to Buy Now?
- B2C Listed the Top Rated Cryptocurrencies for 2023
- Get Early Access to Presales & Private Sales
- KYC Verified & Audited, Public Teams
- Most Voted for Tokens on CoinSniper
- Upcoming Listings on Exchanges, NFT Drops