"Just trying to use the most AI is like maximizing the number of lines of code you write. That's not a good measure of software quality, nor is the number of tokens." Databricks Chief AI Scientist @jefrankle kicked off tonight's DevConnect Meetup at #DataAISummit 2026 by
RESEARCH
-
Consequences of relying on AI for accurate news: MIT study
By
–
The consequences of relying on #AI for accurate news
by Massachusetts Institute of Technology @TechXplore_com Learn more: https://
bit.ly/4e4hpnB #ArtificialIntelligence #MachineLearning #ML -
Codex supports Chrome DevTools protocol for inspecting and modifying websites
By
–
OPENAI 🔥: Codex now supports Chrome DevTools Protocol for browser use. This is a huge superpower that will allow Codex to inspect and modify any website.
— 🚨 AI News | TestingCatalog (@testingcatalog) 16 juin 2026
It is still a very early implementation, but I bet that in several years this will be a default browser capability. If… pic.twitter.com/mvLpRwnUXIOPENAI: Codex now supports the Chrome DevTools protocol for use in the browser. This is a considerable superpower that will allow Codex to inspect and modify any website. This is still a very early implementation, but
-
Cursor accounted for 40-50% of Anthropic revenue early on
By
–
Holy, in its early days, Cursor accounted for roughly 40% to 50% of Anthropic's revenue. And Claude Code was just a reserach project. How quickly everything has changed.
-
Dynamic Linear Attention for Adaptive Token Compression
By
–
"Dynamic Linear Attention" Most long-context linear models compress tokens using fixed blocks or logarithmic schedules, but long texts are not uniform. Stable segments can be summarized, while the
-
Full Analysis of CUA-Bench Benchmark for Computer Use Agents
By
–
Full analysis by @qi_zhengyang
: https://
snorkel.ai/blog/cua-bench
-benchmarking-computer-use-agents-on-professional-software/
… Tasks: https://
cua.ai/cuabench/regis
try/cua-bench-kicad
… -
Google’s 50-page breakdown on AI agent failures and vibe coding shift
By
–
My friend @Saboo_Shubham_
, @AddyOsmani and the team at Google just published a 50-page breakdown on the shift from vibe coding to agentic engineering. It covers the new Software Development Life Cycle with AI Agents. The most interesting takeaway? Most AI agent failures aren't -
Frontier computer-use agent fails on 21 of 25 electrical engineering tasks
By
–
New research: we partnered with @francedot and @ddupont808 at @trycua to stress-test a frontier computer-use agent on real electrical engineering tasks.
— Snorkel AI (@SnorkelAI) 15 juin 2026
25 expert-authored KiCad tasks. 4 passed. 0 build-from-scratch tasks succeeded.
The failure modes are concrete and they point… https://t.co/qHDpHC4jFFNew research: we partnered with @francedot and @ddupont808 at @trycua to stress-test a frontier computer-use agent on real electrical engineering tasks. 25 expert-authored KiCad tasks. 4 passed. 0 build-from-scratch tasks succeeded. The failure modes are concrete and they point
-
When the double descent phenomenon kicks in
By
–
when that double descent kicks in pic.twitter.com/XdxBiLfUdQ
— swyx @aiDotEngineer WF (@swyx) 15 juin 2026when that double descent kicks in
-
Cartesia: a pair of number one models for conversation
By
–
Cartesia did not just release two models, it released the only pair from one company ranked #1 on both ends of a conversation.
— AI Highlight (@AIHighlight) 15 juin 2026
Sonic-3.5 turns text into speech and Ink-2 turns speech into text, faster than anything else streaming.
That is the whole voice loop locked down by… https://t.co/OA6mvEx06nCartesia did not simply release two models; it released the only pair from the same company ranked number one at both ends of a conversation. Sonic-3.5 converts text to speech and Ink-2 converts speech to text, faster than anything else.