Top all-time · #2417
AI21
Company blog · ai21.com/blogai
AI21 is the company blog at ai21.com, writing about ai. Since 2019, 32 different Hacker News readers have linked it 33 times, most often “Jamba: AI21's Groundbreaking SSM-Transformer Model”. No feed was found, so read it on the site itself.
Too few sampled comments (10) to say how often HN recommends it.?
Why HN links it
HN commenters link AI21 Labs' site for its Jurassic-2 LLM studio and API, task-specific business models, and the Jamba model announcement. The comments treat it as a model provider, not as writing to read.
Summarised by Claude from a sample of 10 HN comments that link here. Each comment is marked recommending, critical or neutral (citing without judging); how this works.
AI21's Forward-Deployed Researchers post-train specialized models for your task, creating an AI edge your competitors can’t buy.
Most-linked posts?
- Jamba: AI21's Groundbreaking SSM-Transformer Model5 people · 5 mentions · first linked 2024-03-28
- Jurassic-X: Crossing the Neuro-Symbolic Chasm with the MRKL System4 people · 4 mentions · first linked 2022-04-21
- /3 people · 3 mentions · first linked 2023-03-23
- AI21 Studio and Jurassic-1 Language Models2 people · 2 mentions · first linked 2021-08-11
- /studio2 people · 2 mentions · first linked 2021-12-16
- Multi-scale RAG indexing: why different queries need different chunk sizes2 people · 2 mentions · first linked 2026-02-05
- AI Generated Fiction2 people · 2 mentions · first linked 2019-08-27
- Jamba: A groundbreaking hybrid SSM-Transformer model2 people · 2 mentions · first linked 2024-03-28
What Hacker News says?
My impression is that actually solving this classic RPA problem with AI is exactly the raison d'etre of AI21Labs with their task specific models[1]. They don't have the biggest or best general purpose LLM, but they have an excellent model that's been pre-trained on specific types of business data and also made available for developers using simple APIs & "RPA-style" interfaces. [1] https://www.ai21.com/use-cases
Comment linking /use-cases
on “Computer use, a new Claude 3.5 Sonnet, and Claude 3.5 Haiku”Submitted Jamba 1.5 Open Model Family: The Most Powerful and Efficient Long Context Models as a story · 11 points
HN discussionPlease link to the original post: https://www.ai21.com/blog/announcing-jamba Jamba looks fabulous . Good performance for its size and much more efficient than the available open alternatives. The key idea: One of out of every eight transformer blocks in Jamba applies dot-product attention with quadratic cost, but the other seven out of eight apply a Mamba layer with linear cost. And the entire model is a mixture of…
Comment linking /blog/announcing-jamba
on “Jamba: Production-grade Mamba-based AI model”Submitted Multi-scale RAG indexing: why different queries need different chunk sizes as a story · 8 points
HN discussion@dang this is blogspam for the official post: https://www.ai21.com/blog/announcing-jamba
Comment linking /blog/announcing-jamba
on “Jamba: Production-grade Mamba-based AI model”Submitted Jamba: AI21's Groundbreaking SSM-Transformer Model as a story · 8 points
HN discussion1. https://www.helloforefront.com/ 2. https://www.ai21.com/studio 3. https://huggingface.co/models
Comment linking /studio
on “GPT-3 APIs are now in public beta”Submitted Jurassic-X: Crossing the Neuro-Symbolic Chasm with the MRKL System as a story · 6 points
HN discussionYes and its the same author this time published on the Gradient (the link before was to the personal blog). The Gradient by the way are amazing curators of AI news in general and have one of the better podcasts I am aware of interviewing developers in the trenches. Adding: this resurgence in Mamba in general is also due to some actual sota progress with SSM like the new AI21 lab released this week [1] and likely to…
Comment linking /jamba
on “Mamba Explained”Submitted Released Today – New NLP Model Called Haimke as a story · 6 points
HN discussionAI21 Labs provides access to its Jurassic-2 LLM (and derived models for summarization, grammatical error correction etc.) via a web UI [1] and an API. Disclaimer: I'm a software engineer at AI21. [1] https://www.ai21.com/studio
Comment linking /studio
on “LlamaCloud and LlamaParse”Submitted The Best Reasoning Money (and Time and Compute) Can Buy as a story · 5 points
HN discussionTo me, this looks like a direct competitor to AI21 Labs: https://www.ai21.com/
Comment linking /
on “Mistral Agents”Submitted Introducing Jamba: AI21's Groundbreaking SSM-Transformer Model as a story · 4 points
HN discussionYou can add https://www.ai21.com/ LLMs as well
Comment linking /
on “Show HN: AI Playground by Vercel Labs”Submitted Jamba: A groundbreaking hybrid SSM-Transformer model as a story · 4 points
HN discussionKinda depends on how you define a transformer solving the problem. I feel like you could fine-tune a transformer in the style of https://www.ai21.com/blog/jurassic-x-crossing-the-neuro-symb... to produce steps to solve it, ie "<initial prompt>" => 1. [LLM generate] write a python program to return an integer >500 digits with two different prime factors with >250 digits and return all three 2. [execute code generated…
Comment linking /blog/jurassic-x-crossing-the-neuro-symbolic-chasm-with-the-mrkl-system
on “What a transformer can NOT do”Submitted AI Generated Fiction as a story · 4 points
HN discussionHere are a few... https://www.ai21.com/ https://www.anthropic.com/ https://cohere.ai/ https://huggingface.co/bigscience/bloom https://goose.ai/ https://bard.google.com/ https://ai.facebook.com/blog/large-language-model-llama-meta...
Comment linking /
on “Helicone.ai (YC W23) – Open-Source Logging for OpenAI”Submitted How AI21 manages its GPU fleet with Kueue as a story · 3 points
HN discussionSubmitted Chunk size is query-dependent: a simple multi-scale approach to RAG retrieval as a story · 3 points
HN discussionSubmitted HAIM: An Interpolating Language Model as a story · 3 points
HN discussionSubmitted Lessons from 200k SWE-bench runs as a story · 2 points
HN discussionSubmitted Attention was never enough: Tracing the rise of hybrid LLMs as a story · 2 points
HN discussionSubmitted AI21 Labs concludes largest Turing Test experiment to date as a story · 2 points
HN discussionSubmitted AI21 Studio and Jurassic-1 Language Models as a story · 2 points
HN discussionSubmitted HAIM – Towards Controllable Text Generation as a story · 2 points
HN discussionSubmitted Reaching SOTA on deep research benchmarks by automating agent optimization as a story · 1 points
HN discussionSubmitted Stride and prejudice: How a 32-bit overflow corrupted a CUDA kernel as a story · 1 points
HN discussionSubmitted Jamba: AI21's Groundbreaking SSM-Transformer Model as a story · 1 points
HN discussion
