Our 215th episode with a summary and discussion of last week's big AI news!
Recorded on 07/04/2025
Hosted by <a href='https://twitter.com/andrey_kurenkov'>Andrey Kurenkov</a> and <a href='https://twitter.com/jeremiecharris'>Jeremie Harris</a>. 
Feel free to email us your questions and feedback at <a href='mailto:contact@lastweekinai.com'>contact@lastweekinai.com </a>and/or <a href='mailto:hello@gladstone.ai'>hello@gladstone.ai</a>
Read out our text newsletter and comment on the podcast at <a href='https://lastweekin.ai/'>https://lastweekin.ai/</a>.
In this episode:
<ul>
<li>Cloudflare's new AI data scraper blocking feature, its potential implications, and technical challenges</li>
<li>Meta's aggressive recruitment for its Super Intelligence Labs division is covered, highlighting key hires from OpenAI and other leaders in the field</li>
<li>Anthropic loses significant talent to Cursor, with details on their new economic futures program focusing on AI's impact on the labor market</li>
<li>Notable open-source AI model releases from Baidu and Tencent are also discussed, including their performance metrics and potential applications.</li>
</ul>
Timestamps + Links:
<ul>
<li style="font-weight:400;">(00:00:11) Intro / Banter</li>
<li style="font-weight:400;">(00:01:43) News Preview</li>
</ul>
Tools & Apps
<ul>
<li style="font-weight:400;">(00:02:55) <a href='https://www.nytimes.com/2025/07/01/technology/cloudflare-ai-data.html'>Cloudflare Introduces Default Blocking of A.I. Data Scrapers</a></li>
<li style="font-weight:400;">(00:05:44) <a href='https://www.theverge.com/ai-artificial-intelligence/694531/runway-ai-video-games-generate'>Runway is going to let people generate video games with AI</a></li>
<li style="font-weight:400;">(00:11:24) <a href='https://techcrunch.com/2025/06/30/google-embraces-ai-in-the-classroom-with-new-gemini-tools-for-educators-chatbots-for-students-and-more/'>Google embraces AI in the classroom with new Gemini tools for educators, chatbots for students, and more</a></li>
<li style="font-weight:400;">(00:16:23) <a href='https://www.washingtonpost.com/technology/2025/07/02/ai-note-takers-meetings-bots/'>No one likes meetings. They’re sending their AI note takers instead.</a></li>
<li style="font-weight:400;">(00:18:08) <a href='https://techcrunch.com/2025/06/26/google-launches-doppl-a-new-app-that-lets-you-visualize-how-an-outfit-might-look-on-you/'>Google launches Doppl, a new app that lets you visualize how an outfit might look on you</a></li>
<li style="font-weight:400;">(00:19:14) <a href='https://www.engadget.com/ai/googles-imagen-4-text-to-image-model-promises-significantly-improved-boring-images-111711517.html'>Google's Imagen 4 text-to-image model promises 'significantly improved' boring images</a></li>
</ul>
Applications & Business
<ul>
<li style="font-weight:400;">(00:22:18) <a href='https://www.theverge.com/news/695355/mark-zuckerberg-meta-ai-superintelligence-labs'>Mark Zuckerberg announces his AI ‘superintelligence’ super-group</a></li>
<li style="font-weight:400;">(00:29:35) <a href='https://www.theinformation.com/articles/anthropic-revenue-hits-4-billion-annual-pace-competition-cursor-intensifies?rc=6zqgxt'>Anthropic Revenue Hits $4 Billion Annual Pace as Competition With Cursor Intensifies</a></li>
<li style="font-weight:400;">(00:35:10) <a href='https://techcrunch.com/2025/06/27/as-job-losses-loom-anthropic-launches-program-to-track-ais-economic-fallout/'>As job losses loom, Anthropic launches program to track AI’s economic fallout</a></li>
<li style="font-weight:400;">(00:38:04) <a href='https://finance.yahoo.com/news/openai-says-no-plan-googles-222056614.html'>OpenAI says it has no plan to use Google's in-house chip</a></li>
<li style="font-weight:400;">(00:41:08) <a href='https://www.axios.com/2025/07/01/nvidia-startup-data-center-power'>Nvidia stakes new startup that flips script on data center power</a></li>
<li style="font-weight:400;">(00:44:11) <a href='https://wccftech.com/tsmc-arizona-chips-are-reportedly-being-flown-back-to-taiwan-for-packaging/'>TSMC Arizona Chips Are Reportedly Being Flown Back to Taiwan For Packaging; U.S. Semiconductor Supply Chain Still Remains Dependent on Taiwan</a></li>
</ul>
Projects & Open Source
<ul>
<li style="font-weight:400;">(00:46:57) <a href='https://the-decoder.com/baidu-releases-open-source-model-family-ernie-4-5/'>Baidu releases open source model family ERNIE 4.5</a></li>
<li style="font-weight:400;">(00:51:55) <a href='https://www.marktechpost.com/2025/06/28/tencent-open-sources-hunyuan-a13b-a-13b-active-parameter-moe-model-with-dual-mode-reasoning-and-256k-context/'>Tencent Open Sources Hunyuan-A13B: A 13B Active Parameter MoE Model with Dual-Mode Reasoning and 256K Context</a></li>
<li style="font-weight:400;">(00:57:09) <a href='https://www.marktechpost.com/2025/07/02/together-ai-releases-deepswe-a-fully-open-source-rl-trained-coding-agent-based-on-qwen3-32b-and-achieves-59-on-swebench/'>Together AI Releases DeepSWE: A Fully Open-Source RL-Trained Coding Agent Based on Qwen3-32B and Achieves 59% on SWEBench</a></li>
<li style="font-weight:400;">(01:00:11) <a href='https://arxiv.org/abs/2507.01006'>GLM-4.1V-Thinking: Towards Versatile Multimodal Reasoning with Scalable Reinforcement Learning</a></li>
<li style="font-weight:400;">(01:04:10) <a href='https://arxiv.org/abs/2506.20639v1'>DiffuCoder: Understanding and Improving Masked Diffusion Models for Code Generation</a></li>
</ul>
Research & Advancements
<ul>
<li style="font-weight:400;">(01:06:21) <a href='https://arxiv.org/abs/2503.04412'>Wider or Deeper? Scaling LLM Inference-Time Compute with Adaptive Branching Tree Search</a></li>
<li style="font-weight:400;">(01:13:07) <a href='https://arxiv.org/abs/2506.22419'>The Automated LLM Speedrunning Benchmark: Reproducing NanoGPT Improvements</a></li>
<li style="font-weight:400;">(01:18:04) <a href='https://metr.org/blog/2025-03-19-measuring-ai-ability-to-complete-long-tasks/'>Claude 4 Opus and Sonnet reach 50%-time-horizon point estimates of about 80 and 65 minutes, respectively</a></li>
<li style="font-weight:400;">(01:21:37) <a href='https://arxiv.org/abs/2506.21718'>Performance Prediction for Large Systems via Text-to-Text Regression</a></li>
<li style="font-weight:400;">(01:25:38) <a href='https://arxiv.org/abs/2507.00432v1'>Does Math Reasoning Improve General LLM Capabilities? Understanding Transferability of LLM Reasoning</a></li>
<li style="font-weight:400;">(01:26:33) <a href='https://arxiv.org/abs/2506.07962v1'>Correlated Errors in Large Language Models</a></li>
</ul>
Policy & Safety
<ul>
<li style="font-weight:400;">(01:29:04) <a href='https://forecastingresearch.org/ai-enabled-biorisk'>Forecasting Biosecurity Risks from LLMs</a></li>
<li style="font-weight:400;">(01:36:06) <a href='https://sean-peters-au.github.io/2025/07/02/ai-task-length-horizons-in-offensive-cybersecurity.html'>AI Task Length Horizons in Offensive Cybersecurity</a></li>
<li style="font-weight:400;">(01:42:30) <a href='https://www.obsolete.pub/p/inside-techs-risky-gamble-to-kill'>Inside Tech's Risky Gamble to Kill State AI Regulations for a Decade</a></li>
<li style="font-weight:400;">(01:52:56) <a href='https://www.theguardian.com/technology/2025/jun/27/deepfakes-denmark-copyright-law-artificial-intelligence'>Denmark to tackle deepfakes by giving people copyright to their own features</a></li>
</ul>

Podurama

Home

Trending

Personal

Your Podcasts

New Episodes

My Files

Favourite Episodes

Tags (Folders)

Recently Played

Bookmarks

Followed Playlists

#215 - Runway games, Meta Superintelligence, ERNIE 4.5, Adaptive Tree Search

Last Week in AI

AI Chapters

AI Insights

Reviews