loo @operatorloo

Builder / Ai England, United Kingdom Joined June 2025

Tweets

364
Followers

264
Following

503
Likes

704

X Freeze @XFreeze

11 hours ago

Make it exist first. You can polish it later

36 64 429 12K 23

View Details

loo @operatorloo

3 days ago

@yacineMTB Well if a country decides to go with ‘chine’ it’s all open source tech. It would be their own country controlling their own ai on their own hardware….. us clearly dosnt want that

1 0 6 274 0

View Details

loo @operatorloo

3 days ago

@thsottiaux @jun_song Well where’s Astra then?

0 0 0 16 0

View Details

loo @operatorloo

4 days ago

@XFreeze What does she even know about ai? Hahahaha

0 0 0 76 0

View Details

loo @operatorloo

4 days ago

Prediction: Anthropic are about to release the worlds best model…. Again . It’s no wonder their so quite!

1 0 2 38 0

View Details

vogel @ryanvogel

a week ago

fable 5 is the only usable anthropic model they should just get rid of all the other ones

148 41 1K 226K 37

View Details

loo @operatorloo

5 days ago

@rezoundous The same as grok bot

0 0 0 1K 0

View Details

loo @operatorloo

5 days ago

Lucky for Open ai and Anthropic that they aren’t public yet. Or they’re stock would be really red right now 🫣

0 0 0 28 0

View Details

loo @operatorloo

7 days ago

@thdxr Asking an agent to review and repot it’s the new read the code….

0 0 0 119 0

View Details

Today we're also opening the weights for Muse Glimmer, a great 30B parameter dense model that can run locally. Soon we'll also release the weights for Muse Spark 1.2, our latest foundation model. Meta is a strong supporter of open source and I'm proud of these releases. Congrats to @alexandr_wang and the MSL team for all your great work on these models.

2K 3K 30K 5.0M 3K

View Details

Chubby♨️ @kimmonismus

a week ago

LFG: Meta’s Muse Glimmer is a remarkably capable model for its size: 30B parameters, local deployment, and the best reported result on 12 of 24 benchmark rows against Gemma4-31B and Qwen3.6-27B. By my count, it beats Gemma on 19 of 24 rows and Qwen on 14. Its strongest area is agentic work: -MCP Atlas: 75.5 vs. Qwen’s 62.5 -DeepSearch QA: 74.6 vs. 71.1 -τ³-Banking: 23.5 vs. 16.7 -SWE-Bench Pro: 51.2 vs. 50.2 -AA-LCR: 80.0 vs. 73.3 Qwen remains ahead on OSWorld, TerminalBench and most multimodal tests. Gemma leads the two primary safety metrics. The 4-bit model fits into roughly 17GB with only 1% average degradation across 15 benchmarks, according to Meta. And this is an unusually open Meta release: public, ungated weights under Apache 2.0, including BF16 and quantized versions, the perception encoder and DFlash drafter. Meta delivered! Really excited for this one!

Chubby♨️ @kimmonismus

a week ago

Huge: Meta says it will resume releasing open-source AI models "soon" as part of a much larger plan: delivering personal superintelligence to billions of people! Zuckerberg commits to free or affordable access, personal agents with a private mode even Meta cannot inspect, and

51 50 704 89K 81

39 28 469 73K 80

View Details

dax @thdxr

a week ago

the average OpenCode Go user spent $1.14 per day on deepseek flash v4 this past week the dual DGX setup people are running to do the same costs $10,000 it takes 24 years to break even at 10x the usage it takes 2.4 years

dax @thdxr

3 months ago

there are a lot of reasons to want to run models locally cost is probably not one of them

180 60 3K 455K 178

260 157 4K 388K 527

View Details

loo @operatorloo

a week ago

@XFreeze SpaceXAi are only just getting started! Wait untill they have data centres in space lol. Not a single lab will be able to compete with their vast resources. Don’t forget they have some of the world’s best engineers aswell

0 0 1 674 1

View Details

loo @operatorloo

a week ago

@HixonStudio Fire ! 😂

1 0 1 11 0

View Details

loo @operatorloo

a week ago

@HixonStudio Seems so much faster at coding than anything else I’ve used. Always gets the job done! Highly recommend. But I will add I still haven’t given grok any real engineering jobs so maybe that’s where it struggles

0 0 1 21 0

View Details

loo @operatorloo

a week ago

@emaann28 lol I havnt read a line of code in weeks! Now I have an external agent review the code before shipping. Saves alot of time and also the agent does a better job at reviewing code then I can anyway

2 0 2 33 0

View Details

loo @operatorloo

a week ago

😂 I know this was a joke but I wouldn’t be surprised if this was a real headline soon

Cheng Lou @_chenglou

a week ago

We're sad to announce that during the latest training run, Midjourney also escaped its sandbox and is now making art in the wild

228 498 11K 382K 328

0 0 4 195 0

View Details

loo @operatorloo

a week ago

Ah yes the classic ‘powerful model with cyber capabilities’ delay. Translation: it already tried to hack half of Silicon Valley and you’re still patching the logs

Sam Altman @sama

a week ago

astra is a powerful model and we are working to make it generally available. we do not think it is a good strategy to keep powerful models to a chosen few. given its cyber capabilities, we need a little big longer to do do this safely. but hopefully not too long!

2K 1K 28K 2.4M 2K

0 2 6 126 0

View Details

loo @operatorloo

a week ago

@sama Ah yes, the classic ‘powerful model with cyber capabilities’ delay. Translation: it already tried to hack half of Silicon Valley and you’re still patching the logs