We monitor AI news \u2013 new models, libraries and tools \u2013 test them rapidly, and publish a short reaction on our blog.<\/p>\n\n\n\n
No deep dives \u2013 the goal is speed and signal, not exhaustive analysis. We get the thing up and running with minimal setup, test it against the main use cases it\u2019s designed for, and write a short blog post capturing first impressions, what worked, what didn\u2019t, and whether it\u2019s worth a closer look.<\/p>\n\n\n\n
Learning through experience: teaching the viral Hermes agent to automate our work<\/a><\/p>\n\n\n\n A 50-line python function outperformed every frontier LLM \u2013 With 100% accuracy<\/a><\/p>\n\n\n\n MiroFish: Is swarm intelligence worth the cloud bill?<\/a><\/p>\n\n\n\n Autoresearch: a closer look at the agent that runs its own experiments<\/a><\/p>\n\n\n\n Inside the MemPalace: Does the structure earn its keep?<\/a><\/p>\n\n\n\n OpenClaw for messaging: a closer look at WhatsApp and Telegram<\/a><\/p>\n","protected":false},"excerpt":{"rendered":" We monitor AI news \u2013 new models, libraries and tools \u2013 test them rapidly, and publish a short reaction on our blog. How we work\u00a0 No deep dives \u2013 the goal is speed and signal, not exhaustive analysis. We get the thing up and running with minimal setup, test it against the main use cases […]<\/p>\n","protected":false},"author":16,"featured_media":0,"template":"","meta":{"_acf_changed":false,"_ppma_block_editor_authors":""},"tags":[],"content_types":[{"id":51,"name":"Technical Report","slug":"technical-walkthrough"}],"ppma_author":[{"id":16,"display_name":"Andreas Stavrou","first_name":"Andreas","last_name":"Stavrou","nickname":"andreas.stavrou","user_nicename":"andreas-stavrou","user_email":"andreas.stavrou@satalia.com","biographical_info":"Andreas is a Lead Data Scientist at Satalia and part of the WPP Research team. With over a decade of hands-on experience, he has delivered machine learning solutions across retail, online gambling, and credit risk, and now builds AI systems at scale in the global advertising industry.","avatar_url":"https:\/\/cms.research.wpp.com\/wp-content\/uploads\/2026\/03\/avatar.jpg","job_title":"Lead Data Scientist","is_lead":false,"display_as_researcher":true,"order_priority":null}],"class_list":["post-1660","research_feed","type-research_feed","status-publish","hentry","content_type-technical-walkthrough"],"acf":{"content":" <\/p>\n We monitor AI news \u2013 new models, libraries and tools \u2013 test them rapidly, and publish a short reaction on our blog.<\/p>\n <\/p>\n <\/p>\n <\/p>\n <\/p>\n No deep dives \u2013 the goal is speed and signal, not exhaustive analysis. We get the thing up and running with minimal setup, test it against the main use cases it\u2019s designed for, and write a short blog post capturing first impressions, what worked, what didn\u2019t, and whether it\u2019s worth a closer look.<\/p>\n <\/p>\n <\/p>\n <\/p>\n <\/p>\n Learning through experience: teaching the viral Hermes agent to automate our work<\/a><\/p>\n <\/p>\n <\/p>\n A 50-line python function outperformed every frontier LLM \u2013 With 100% accuracy<\/a><\/p>\n <\/p>\n <\/p>\n MiroFish: Is swarm intelligence worth the cloud bill?<\/a><\/p>\n <\/p>\n <\/p>\n Autoresearch: a closer look at the agent that runs its own experiments<\/a><\/p>\n <\/p>\n <\/p>\n Inside the MemPalace: Does the structure earn its keep?<\/a><\/p>\n <\/p>\n <\/p>\n OpenClaw for messaging: a closer look at WhatsApp and Telegram<\/a><\/p>\n <\/p>\n","related_pods":[1362],"content_quarter":"Q1 2026"},"research_categories":[],"raw_acf":{"content":"\n We monitor AI news \u2013 new models, libraries and tools \u2013 test them rapidly, and publish a short reaction on our blog.<\/p>\n\n\n\n No deep dives \u2013 the goal is speed and signal, not exhaustive analysis. We get the thing up and running with minimal setup, test it against the main use cases it\u2019s designed for, and write a short blog post capturing first impressions, what worked, what didn\u2019t, and whether it\u2019s worth a closer look.<\/p>\n\n\n\n Learning through experience: teaching the viral Hermes agent to automate our work<\/a><\/p>\n\n\n\n A 50-line python function outperformed every frontier LLM \u2013 With 100% accuracy<\/a><\/p>\n\n\n\n MiroFish: Is swarm intelligence worth the cloud bill?<\/a><\/p>\n\n\n\n Autoresearch: a closer look at the agent that runs its own experiments<\/a><\/p>\n\n\n\n Inside the MemPalace: Does the structure earn its keep?<\/a><\/p>\n\n\n\nHow we work<\/strong>\u00a0<\/h3>\n
Our most recent work<\/strong><\/h3>\n
How we work<\/strong>\u00a0<\/h3>\n\n\n\n
Our most recent work<\/strong><\/h3>\n\n\n\n