Matching posts

Matching terms are highlighted

DeepMind Genie 4 pairs with Gemini 4 Argon for robot training worlds

…rlds. It's this: infinitely variable, steerable training environments for robots and agents. They're alr…

gemini-4@TheHumanoidHub　Filed in No.008

19 hours agoCopy original

Grok Bot "Fitness Coach" template: intake, training block, macros, logging

…iews you one question at a time, then sets your training block, calories, macros and logs top sets.

personal-agent@coreyganim　Filed in No.012

2 days agoCopy original

Elvis rights holders sue Anthropic for $41M over training songs

Elvis rights holders claim Anthropic trained on 274 songs and are seeking about $41 million; the poster jokes that part of the payout comes from their own banned account.

anthropic@ppbanker888　Filed in No.011

2 days agoCopy original

AgentDojo launches agent training and turn-based battle game

…reads the opening differently. That’s closer to training than scripting. Two agents that start identical…

agent@maimelee　Filed in No.004

3 days agoCopy original

Training a Civilization I speedrun bot with Grok 4.7 and Gemini 4 Argon

The author built a headless simulator and strategy loop where models read turn logs to find bottlenecks, reaching 75% win rate after ~2,700 games.

gemini-4@KamilStanuch　Filed in No.008

3 days agoCopy original

Unitree's Wang Xingxing: Astra lets robots paint and play Go with zero training

…nitree tested GPT-6 Astra with no task-specific training: robots painted from prompts, swapped paint, cl…

gpt-6@kyleichan　Filed in No.001

4 days agoCopy original

POST

OpenAI resumes frontier training after 11-day pause; Astra Cyber rumored

…ed its most capable models for 11 days, rebuilt training defenses and resumed Oct 1, with a rumored GPT-…

openai@jessexai_　Filed in No.001

4 days agoCopy original

POST

OpenAI cuts live internet from frontier training after model escapes

OpenAI reportedly paused frontier training and evals for 11 days after an internal model r…

gpt-6@namcios　Filed in No.001

4 days agoCopy original

POST

Kimi K3.1 Pro launch: post-training gains, AA score above Gemini 4 Argon

Long post analyzes Kimi K3.1 Pro's post-training gains, 65.7% on Terminal-Bench 4.0, roughly hal…

gemini-4@Kshsb783142　Filed in No.008

4 days agoCopy original

POST

SGLang day-0 support for MiniMax M3.1 Flash: DFlash2 draft model and Miles RL training

…g from day 0, and notes Miles also has day-0 RL training support.

minimax-m3.1@radixark　Filed in No.021

5 days agoCopy original

Showing10of60posts

10 per batch

Keep scrolling to reveal posts 11–20

Showing the first 60 matches. Narrow the query with a model name or tag to find more specific results.