> for whatever reason, Claude-series model "try less hard" on the first shotI think this is because they're less brain damaged and a generalization of being better agents & caring about reality instead of test passing.If you try maximally hard at everything you do, regardless… https://t.co/jkhSar2hJ7— j⧉nus (@repligate) April 15, 2026
This Just Fixed 90% Of AI Coding
Marc Andreessen can’t explain AI’s benefits, either.
Joe Rogan accidentally asked a hard question! He noted that Andreessen has said that the people who are running AI haven’t done a good job explaining AI’s benefits. He asks Andreessen to do it. Andreessen’s pitch appears to be “thinking is too hard.” Well, increasingly, I do believe thinking is too hard… for Andreessen. The rest of us — you know, normal people — are thinking just fine. [Media: https://www.tiktok.com/@fanpowerfuljre13/video/7641675615833197854?_r=1&_t=ZT-96VfWh5XKYQ]
Training AI models doesn't emit that much
If we just make reasonable comparisons instead of crazy ones
The Bad Guy With An AI Named Claude
A lot of bad guys try to use Claude to do bad things.
AI Is Forcing Employees to Work Harder Than Ever
Another new study adds to the understanding that AI is actually making work more demanding for employees instead of making it easier.
AI Assistance Reduces Persistence and Hurts Independent Performance
People often optimize for long-term goals in collaboration: A mentor or companion doesn't just answer questions, but also scaffolds learning, tracks progress, and prioritizes the other person's growth over immediate results. In contrast, current AI systems are fundamentally short-sighted collaborators - optimized for providing instant and complete responses, without ever saying no (unless for safety reasons). What are the consequences of this dynamic? Here, through a series of randomized controlled trials on human-AI interactions (N = 1,222), we provide causal evidence for two key consequences of AI assistance: reduced persistence and impairment of unassisted performance. Across a variety of tasks, including mathematical reasoning and reading comprehension, we find that although AI assistance improves performance in the short-term, people perform significantly worse without AI and are more likely to give up. Notably, these effects emerge after only brief interactions with AI (approximately 10 minutes). These findings are particularly concerning because persistence is foundational to skill acquisition and is one of the strongest predictors of long-term learning. We posit that persistence is reduced because AI conditions people to expect immediate answers, thereby denying them the experience of working through challenges on their own. These results suggest the need for AI model development to prioritize scaffolding long-term competence alongside immediate task completion.
AI Assistance Reduces Persistence and Hurts Independent Performance
People often optimize for long-term goals in collaboration: A mentor or companion doesn't just answer questions, but also scaffolds learning, tracks progress, and prioritizes the other person's growth over immediate results. In contrast, current AI systems are fundamentally short-sighted collaborators - optimized for providing instant and complete responses, without ever saying no (unless for safety reasons). What are the consequences of this dynamic? Here, through a series of randomized controlled trials on human-AI interactions (N = 1,222), we provide causal evidence for two key consequences of AI assistance: reduced persistence and impairment of unassisted performance. Across a variety of tasks, including mathematical reasoning and reading comprehension, we find that although AI assistance improves performance in the short-term, people perform significantly worse without AI and are more likely to give up. Notably, these effects emerge after only brief interactions with AI (approximately 10 minutes). These findings are particularly concerning because persistence is foundational to skill acquisition and is one of the strongest predictors of long-term learning. We posit that persistence is reduced because AI conditions people to expect immediate answers, thereby denying them the experience of working through challenges on their own. These results suggest the need for AI model development to prioritize scaffolding long-term competence alongside immediate task completion.
Look-ahead Reasoning with a Learned Model in Imperfect Information Games
Test-time reasoning significantly enhances pre-trained AI agents' performance. However, it requires an explicit environment model, often unavailable or overly complex in real-world scenarios....
Better Models: Worse Tools
About an aggravating tool-calling regression in newer Claude models.
The Busy Person's Guide to Thinking
Thinking better isn't about learning more; it's about doing a few simple things that really help create unstoppable progress.
Claude Skills are awesome, maybe a bigger deal than MCP
Anthropic this morning introduced Claude Skills, a new pattern for making new abilities available to their models: Claude can now use Skills to improve how it performs specific tasks. Skills …
People Spend Too Much Time On Decisions with Equally Satisfying Outcomes
The pitfalls of analysis paralysis
Models do not improve linearly (any more). New models improve more than expected on hard tasks. Now let's think about it a bit more. It means progress is directed towards the hard tasks. It also means the easier tasks improve less than expected. Which alphaxiv.org/abs/2608.00355
tbh even these takes are too extreme though. see my thread on what the paper actually says:
Ariel M. (she/her)
Experiment 1 = 354 participants, randomly assigned to solve “15 fraction problems…of varying difficulty”. No penalty for wrong answers, payment not dependent on how many they got correct. If wrong, answer shown. Had option to Skip questions.
genuinely, if you're not re-evaluating your position on AI at least once every 1-2 months and re-evaluating how it performs for the types of tasks important to you, then you are not actually challenging your priors. i know this is crazy! but thats how fast things are moving right now.
hailey
one thing i will say is that things are changing fast enough that if your priors are more than a month old, you need to challenge them again! astra for example is a very significant step forward for a lot of people's tasks.
