Large language models (LLMs) have evolved into agentic systems capable of autonomous tool use and multi-step reasoning for complex problem-solving. However, post-training approaches building upon general-purpose foundation models consistently underperform in agentic tasks, particularly in open-source implementations. We identify the root cause: the absence of robust agentic foundation models forces models during post-training to simultaneously learn diverse agentic behaviors while aligning them to expert demonstrations, thereby creating fundamental optimization tensions. To this end, we are the first to propose incorporating Agentic Continual Pre-training (Agentic CPT) into the deep research agents training pipeline to build powerful agentic foundational models. Based on this approach, we develop a deep research agent model named AgentFounder. We evaluate our AgentFounder-30B on 10 benchmarks and achieve state-of-the-art performance while retains strong tool-use ability, notably 39.9% on BrowseComp-en, 43.3% on BrowseComp-zh, and 31.5% Pass@1 on HLE.

An employee cleans the fuselage of a Gulfstream G-650 jet on the second day of the Paris Air Show in Paris, France, on Tuesday, June 18, 2013. The 50th International Paris Air Show is the world's largest aviation and space industry show, and takes Over the last couple of years, the concept of aerial ridesharing, aided by the tech back-end of the obligatory mobile app, has been a pre-occupation of several startups. The goal has consistently been to take the not necessarily affordable concept of flying privately and make it affordable. There have been many attempts in the market as the trend continues to build, but clearly no strongholds just yet. From Jetsmarter to Fly by Airthereum and several others, most have yet to reach the wheels up level much less remain in in the skies.

At the 2019 Paris Air Show, Dassault Aviation showed a mockup of the FCAS sixth generation fighter jet for the first time publicly. A new partnership between ANSYS and Airbus Defense and Space will develop a new artificial intelligence design tool to create the embedded flight control software for Europe's Future Combat Air System (FCAS). FCAS is a next-generation air combat development program involving France, Germany and now Spain to develop a system of fully automated remote air platforms and sixth-generation fighters that will replace the current generation of Eurofighter and Rafale jets operated by those three countries. Dassault and Airbus are the lead prime manufacturers for the FCAS program. A mockup of the future FCAS stealth fighter concept was shown publicly for the first time during the 2019 Paris Air Show.

There are flying cars and Concorde's would-be supersonic successor, a company offering to deliver cargo to the Moon - for a mere $1.2 million per kilogram - and the latest in funky futuristic aviation ideas, both big and small. No doubt about it: the Paris Air Show is an aerospace geek's paradise. But with everything from the smallest drones to the largest passenger jets on display, it's tough to sift through it all. So here's a guide to some of the cool things that caught our eye this week. Visitors looks at the flying car Pegasus 1, built by French entrepreneur Jerome Dauffy at Paris Air Show, in Le Bourget, east of Paris, France, Tuesday, June 20, 2017 in Paris.

The head of the U.S. Federal Aviation Administration says the boom in drone popularity is'a huge game changer' similar'to powered flight or jet engines.' In an Associated Press interview Tuesday, Michael Huerta said'the growth of this industry and how it's evolving is something that all of us in aviation need to pay a lot of attention to.' Huerta spoke on the sidelines of the Paris Air Show, where drones of all kinds were being displayed. Michael Herta, the head of the U.S. Federal Aviation Administration, speaks at the Paris Air Show on Tuesday, June 20, 2017. In an Associated Press interview, Huerta called the rapid development of the drone industry'a huge game-changer' for aviation. For the industry, their popularity begs a vital question: how do you make sure that swarms of new machines don't endanger each other, other users of the skies and people on the ground as they do everything from patrolling traffic blackspots to, possibly, delivering your burger and fries?