Co-located with ICFP 2024 in Milan this year is the Functional Software Architecture (FUNARCH) workshop. Come and participate in discussions about the construction and structure of large and long-lived software projects that are implemented in functional languages, and listen to some talks that touch on topics like feature engineering; applying continuous formal methods to the proof-of-stake cryptocurrency Cardano; and bidirectional data transformations. @ACMFUNARCH functional-architecture.org/…

Please consider submitting a paper to The Second ACM SIGPLAN Workshop on Functional Software Architecture - FP in the Large if you have stories to share about methods for the construction of large and long-lived software projects that are implemented in functional languages and released to real users. The workshop occurs on 6 September in Milan, co-located with ICFP. The deadline for submission is the 3rd of June 2024. functional-architecture.org/…

We’ve developed an LLM chat service that runs at breathtaking speed compared to some other LLM chat services you may have used. Please give it a try via groq.com. The LLM chat service is built using our special purpose chips (LPUs) for accelerating machine learning algorithms which are designed to compose and scale with LEGO-like snapability, with determinism being our glue; of our compiler that maps open source foundational machine learning models onto our compute architecture using what appears to be a sequence of lightening Jujitsu moves; and of our cloud and operations teams that stood up and maintain the datacenter hardware that serves a foundational model robustly at scale, speed and low latency. The specific model we serve is Llama2, 70B 4k sequence, created by Meta AI. My day to day life at Groq involves working deep in the trenches on things like SystemVerilog RTL for power management circuits; a #Haskell DSL for structured cell-level digital design; Haskell DSLs for generating machine learning and HPC workloads; and front end MLIR compiler work for wrangling tensors and ops from ONNX, StableHLO, #JAX etc. It is easy to get lost in the details and loose sight of the big picture of what the whole company is striving towards. However, seeing this LLM demo in action makes concrete in a visceral way what the collaboration of many people’s hard work has achieved. I hope this gives you a sense of what I work on when I am not cooking, and share in our excitement. This is an alpha version, more is coming in Q1 2024. #betterongroq #groqspeed #llm #genai