• AI Time to Impact
  • Posts
  • . . AI: Claude's Big Launch; Indian Government Clamp Down (3.4.24)

. . AI: Claude's Big Launch; Indian Government Clamp Down (3.4.24)

Anthropic, AI Bias, Baseten ML Opps

Author

Marshall Kirkpatrick
March 04, 2024

Friends, just three stories dominate the AI community conversation today. Powerful new versions of Claude, a rather shocking AI policy decision in India, and a big B round of funding for ML Ops startup Baseten.

What do these systems all have in common? They are about the smooth, reliable, safe operation of artificial intelligence. But, the new Claude and I just had a little chat about deterministic, mechanical systems needing maintenance and non-deterministic, emergent systems needing management: it recommended the catchy phrase “grease the gears, tame the entropy.” But when I suggested I liked “ride the entropy” better, Claude said it did too.

Now here’s today’s AI news,

Marshall Kirkpatrick, Editor

First impacted: Claude users, AI developers
Time to impact: Short

Anthropic has launched its Claude 3 model series which will be available for use via claude.ai and the Claude API. In order of increasing capability, the series includes Claude 3 Haiku, Claude 3 Sonnet, and Claude 3 Opus. Anthropic claims Claude 3 Opus outperforms GPT-4 and Gemini 1.0 Ultra in various benchmarks across math, reasoning, and code. We've found Claude to be the best system for understanding PDFs and Sonnet continues to do a good job with that. One test of its OCR capabilities returned an analysis of a handwritten exercise log filled with hallucinated workouts and workout buddies, though.

  • According to Anthropic, Claude 3 Opus boasts an impressive capability to achieve nearly flawless recall, surpassing 99% accuracy. The company claims it can even detect instances where a test sentence appears to have been humanly inserted into the original text.

  • AWS has revealed intentions to introduce the Claude 3 AI models on its Amazon Bedrock platform, with Claude 3 Sonnet currently accessible.

Indian Ministry Issues AI Advisory to Curb Bias, Ensure Electoral Integrity

First impacted: Indian internet users, Companies using AI models and generative AI software in India
Time to impact: Long

India's Ministry of Electronics and Information Technology has issued an advisory requiring companies to obtain government approval before deploying AI models, LLMs, and generative AI software on the Indian internet, according to The Economic Times. The advisory calls for companies to ensure their AI systems do not breed bias, discrimination, or undermine electoral processes. The policy was widely critiqued in international AI circles; Abacus AI CEO Bindu Reddy, for example, posted a representative post on X, starting with the words "India just kissed its future goodbye!" [MeitY approval must for companies to roll out AI, generative AI models] Explore more of our coverage of: AI Regulation, Indian Tech Policy, Bias in AI. Share this story by email

First impacted: Business owners, IT professionals
Time to impact: Medium

Cloud-based machine learning infrastructure provider, Baseten, says it has raised $40 million in a recent funding round, led by IVP and Spark. "For most businesses," the company says, "it is still very challenging to run large models in production. Every part of the problem is difficult, cumbersome, and expensive: acquiring compute, getting models to run fast, scaling to meet demand, observability, CI/CD, cost optimization — the list goes on. When handling it all on their own, most teams spend as much time running models in production as they do focusing on the core product experiences they are building." [Announcing our Series B] Explore more of our coverage of: Machine Learning, Cloud Infrastructure, Startup Funding. Share this story by email