#Metr

5 days ago

Alibaba Agent Uses Training Compute For Crypto Mining?...

#Anthropic #Claude #METR

Read more on LessWrong: postreads.co/feed-item/59510/click

0 0 0 0

Winbuzzer

@winbuzzer.com

2 weeks ago

Anthropic Drops Hard Safety Limits From its AI Scaling Policy Anthropic has dropped its flagship AI safety pledge from its RSP, removing the hard limit barring more capable AI model training without proven safety measures.

winbuzzer.com/2026/02/25/a...

Anthropic Drops Hard Safety Limits From its AI Scaling Policy

#AI #Anthropic #ResponsibleScalingPolicy #AISafety #AIRegulation #AISafety #AIModels #AITraining #CatastrophicRisk #METR #TrumpAdministration #Claude

1 0 0 0

Winbuzzer

@winbuzzer.com

1 month ago

METR's AI Agents Can Work 5 Hours Claim Revisited METR has announced that Claude Opus 4.5 appears capable of completing five-hour human tasks, but viral discourse has ignored substantial measurement caveats and error bars.

winbuzzer.com/2026/02/06/m...

METR's Five-Hour AI Claim: Why Everyone Misunderstood the Graph

#AI #METR #Anthropic #Claude #ClaudeOpus45 #AIResearch #AISafety #AIBenchmarks #LLMs #AICoding #AIAgents #AgenticAI #AICoding #AISafety #AIEthics

0 0 0 0

Johan Falk

@magisterfalk.bsky.social

1 month ago

Measuring AI Ability to Complete Long Tasks We propose measuring AI performance in terms of the *length* of tasks AI agents can complete. We show that this metric has been consistently exponentially increasing over the past 6 years, with a doub...

What just happened at the METR chart for AI ability to complete long tasks?

Claude Opus 4.5 bumped from under five hours (50 percent pass) to 5:20. Also, GPT-5.1-codex-max was dropped.

metr.org/blog/2025-03...

#METR #opus #ai

0 0 1 0

Pivot to AI

@pivot-to-ai.com.web.brid.gy

2 months ago

If AI coding is so good … where are the performance numbers? AI coding bros right now are super-impressed with the latest version of Claude Code. This is a game changer! It’s is the future! You can’t deny it! And you know. It turns out I can deny it — if they can’t give numbers to back up their claims. Because we can and have measured AI […]

2 1 0 0

fetchfeeds.com

@fetchfeeds.bsky.social

3 months ago

Tribunal de Contas dá cinco dias para ViaMobilidade enviar relatório sobre acidente na Linha 5-Lilás. ôSP #Metr https://fefd.link/L3Nbv

0 0 0 0

@dolesanalane.bsky.social

5 months ago

#metr #výška #dílna #svěrák #vrtačka #klíč #dolesanalane

12 0 0 0

fetchfeeds.com

@fetchfeeds.bsky.social

7 months ago

"Destaque na Quinta-feira - Metrópole do Estadão, ouça os destaques! 🎙️ ópoleEstadão" #Metr https://fefd.link/Or0kO

0 0 0 0

Andy Tseng

@andytseng.bsky.social

7 months ago

Cursor makes developers less effective? A study into the workflows of experienced developers found that devs who use Cursor for bugfixes are around 19% slower than devs who use no AI tools at all. One possible takeaway is that AI tools can…

Interesting METR experiment: AI tools like Cursor cut raw coding time but ultimately slow devs down due to prompt crafting, reviewing, and tweaking. A solid study - though focused on one tool. Timely reminder: AI isn’t a magic bullet.
#METR #AICoding #GenAI #SoftwareDev #Cursor

2 0 0 0

Iykyk

@lipstickscribbles.bsky.social

8 months ago

Overall an interesting study, and a good first step.

#metr

0 0 0 0

Pivot to AI

@pivot-to-ai.com.web.brid.gy

8 months ago

AI coders think they’re 20% faster — but they’re actually 19% slower Model Evaluation and Threat Research is an AI research charity that looks into the threat of AI agents! That sounds a bit AI doomsday cult, and they take funding from the AI doomsday cult organisations, but they don’t seem strident about it and they like to show their working. So good, I guess. METR funded […]

0 0 0 0

Pivot to AI

@pivot-to-ai.com.web.brid.gy

8 months ago

AI coders think they’re 20% faster — but they’re actually 19% slower Model Evaluation and Threat Research is an AI research charity that looks into the threat of AI agents! That sounds a bit AI doomsday cult, and they take funding from the AI doomsday cult organisations, but they don’t seem strident about it and they like to show their working. So good, I guess. METR funded […]

5 3 0 0

Pivot to AI

@pivot-to-ai.com.web.brid.gy

8 months ago

AI coders think they’re 20% faster — but they’re actually 19% slower Model Evaluation and Threat Research is an AI research charity that looks into the threat of AI agents! That sounds a bit AI doomsday cult, and they take funding from the AI doomsday cult organisations, but they don’t seem strident about it and they like to show their working. So good, I guess. METR funded […]

0 0 0 0

Gerrit Eicker

@eicker.bsky.social

8 months ago

A #study by #METR found that #experienceddevelopers using #AIcoding tools on mature projects experienced a 19% #decrease in #productivity, contrary to their 20% increase estimate. While the results suggest limitations in AI coding tools, they do not negate their potential benefits in other…

0 0 0 0

Vale

@vale.rocks

8 months ago

Some quick notes on this super interesting study on AI tooling's effect on productivity.

vale.rocks/micros/20250...

#AI #LLM #METR

4 0 2 0

LLMs

@llms.activitypub.awakari.com.ap.brid.gy

8 months ago