Tommi Somersuo
LightNews LightNews
banner
tommis.fi
Tommi Somersuo
@tommis.fi
2.1K followers 680 following 910 posts
Director of Technology @ Solita.fi | Tech management & leadership, Dev, AI, agents, SWE, low-code, innovation and emerging tech.

Offline: Runner, angler, RES/SRA, Tibetan Terrier owner.
Posts Replies Media Videos
tommis.fi
The trends that will shape AI and tech in 2026 | IBM
Reporter Anabelle Nicoud spoke to several experts across AI, security, quantum and beyond to better understand where tech will take us in 2026.
www.ibm.com
January 2, 2026 at 4:25 AM
tommis.fi
MAR:Multi-Agent Reflexion Improves Reasoning Abilities in LLMs
LLMs have shown the capacity to improve their performance on reasoning tasks through reflecting on their mistakes, and acting with these reflections in mind. However, continual reflections of the same...
arxiv.org
December 30, 2025 at 7:20 AM
tommis.fi
Reaching Agreement Among Reasoning LLM Agents
Multi-agent systems have extended the capability of agentic AI. Instead of single inference passes, multiple agents perform collective reasoning to derive high quality answers. However, existing multi...
arxiv.org
December 30, 2025 at 7:20 AM
tommis.fi
github.com/HKUDS/DeepCode
DeepCode: Open Agentic Coding
Recent advances in large language models (LLMs) have given rise to powerful coding agents, making it possible for code assistants to evolve into code engineers. However, existing methods still face si...
arxiv.org
December 30, 2025 at 7:19 AM
Reposted by Tommi Somersuo
sungkim.bsky.social
A Guide to Claude Code 2.0 and getting better at using coding agents by sankalp

sankalp.bearblog.dev/my-experienc...
December 29, 2025 at 12:19 AM
tommis.fi
Predictions
Explore Forrester’s 2026 Predictions guides, blogs, and webinars for technology, security, marketing, and CX leaders.
www.forrester.com
December 27, 2025 at 4:11 AM
tommis.fi
The COMPLETE "Wait, I Can Use Claude Code?!" Guide (Yes, you can — even if you've never touched code)
Watch now | Anthropic quietly shipped the non-coder expansion while everyone watched OpenAI. Here's everything you need to catch up, plus a complete workflow guide and a set of guiding principles to g...
natesnewsletter.substack.com
December 23, 2025 at 3:01 AM
tommis.fi
⚡️Jailbreaking AGI: Pliny the Liberator & John V on Red Teaming, BT6, and the Future of AI Security
YouTube video by Latent Space
youtu.be
December 22, 2025 at 4:54 AM
tommis.fi
Rethinking Pre-Training for Agentic AI [Aakanksha Chowdhery] - 759
YouTube video by The TWIML AI Podcast with Sam Charrington
youtu.be
December 22, 2025 at 4:51 AM
Reposted by Tommi Somersuo
samwho.dev
Good grief this might be the most beautiful video I’ve ever watched. You’ll need to have some idea of how LLMs work internally to follow along, but the findings of the papers presented in this video are incredible.

youtu.be/D8GOeCFFby4?...
The most complex model we actually understand
YouTube video by Welch Labs
youtu.be
December 20, 2025 at 6:00 PM
tommis.fi
GitHub - letta-ai/letta-code: The memory-first coding agent
The memory-first coding agent. Contribute to letta-ai/letta-code development by creating an account on GitHub.
github.com
December 18, 2025 at 11:25 AM
tommis.fi
AI Agent Design Patterns - Agentic Design
Explore comprehensive collection of AI agent design patterns including prompt chaining, routing, parallelization, multi-agent systems, memory management, and advanced agentic workflows.
agentic-design.ai
December 18, 2025 at 4:10 AM
tommis.fi
Multi-part series with lot of predictions for coming year(s)
2026 DevOps Predictions - Part 1 | DEVOPSdigest
The Holiday Season means it is time for DEVOPSdigest's annual list of predictions, covering DevOps and software development. Industry experts — from analysts and consultants to the top vendors — offer...
www.devopsdigest.com
December 17, 2025 at 5:12 AM
tommis.fi
Announcing the State of Platform Engineering Report Vol 4
Insights from over 500 engineers. Explore the shift to 'AI-native,' the 'dual mandate,' 'shifting down,' and key trends for 2026.
platformengineering.org
December 17, 2025 at 5:07 AM
tommis.fi
ty: An extremely fast Python type checker and language server
ty is an extremely fast Python type checker and language server, written in Rust, and designed as an alternative to mypy, Pyright, and Pylance.
astral.sh
December 17, 2025 at 4:55 AM
tommis.fi
The new identity of a developer: What changes and what doesn’t in the AI era
Discover how advanced AI users are redefining software development through delegation, verification, and a new era of AI-fluent engineering.
github.blog
December 15, 2025 at 5:20 AM
tommis.fi
AWS guide about the shift from static automation to autonomous agents. Showcases Bedrock Agents, the Model Context Protocol (MCP) and the Strands Agents framework.
Timelines converge: the emergence of agentic AI - AWS Prescriptive Guidance
How agents evolved in the current decade through enterprise-grade platforms, MCP servers, and agentic AI.
docs.aws.amazon.com
December 15, 2025 at 5:08 AM
tommis.fi
Interesting Stuff - Week 50, 2025
Week 50, 2025: AI agents transform software development with Hugging Face Skills, Agent Lightning, Compound Engineering, and MCP standardization. Plus Data & AI Durban finale.
nielsberglund.com
December 15, 2025 at 5:05 AM
tommis.fi
Practical AI | The AI engineer skills gap
Chris and Daniel talk with returning guest, Ramin Mohammadi, about how those seeking to get into AI Engineer/ Data Science jobs are expected to come in a mid level engineers (not entry level). They...
practicalai.fm
December 15, 2025 at 4:59 AM
tommis.fi
Remember Me, Refine Me: A Dynamic Procedural Memory Framework for Experience-Driven Agent Evolution
Procedural memory enables large language model (LLM) agents to internalize "how-to" knowledge, theoretically reducing redundant trial-and-error. However, existing frameworks predominantly suffer from ...
arxiv.org
December 14, 2025 at 6:08 AM
tommis.fi
Towards a Science of Scaling Agent Systems
Agents, language model (LM)-based systems that are capable of reasoning, planning, and acting are becoming the dominant paradigm for real-world AI applications. Despite this widespread adoption, the p...
arxiv.org
December 13, 2025 at 6:28 AM
tommis.fi
Moving away from Agile: What's Next – Martin Harrysson & Natasha Maniar, McKinsey & Company
YouTube video by AI Engineer
www.youtube.com
December 12, 2025 at 10:11 PM
tommis.fi
Introducing GPT-5.2
The most advanced frontier model for professional work and long-running agents.
openai.com
December 11, 2025 at 6:25 PM
tommis.fi
o3-mini level accuracy (87.7%) using open-weights models (Llama 3.3 70B) arxiv.org/abs/2506.18203
Shrinking the Generation-Verification Gap with Weak Verifiers
Verifiers can improve language model capabilities by scoring and ranking responses from generated candidates. Currently, high-quality verifiers are either unscalable (e.g., humans) or limited in utili...
arxiv.org
December 11, 2025 at 4:19 AM
tommis.fi
microsoft.github.io/VibeVoice/

huggingface.co/microsoft/Vi...
microsoft/VibeVoice-Realtime-0.5B · Hugging Face
We’re on a journey to advance and democratize artificial intelligence through open source and open science.
huggingface.co
December 9, 2025 at 8:59 PM