acl (@acleverthought) header
acl (@acleverthought) profile picture
Joined April 2014
Photo by acl (@acleverthought)Photo by acl (@acleverthought)Photo by acl (@acleverthought)Photo by acl (@acleverthought)Photo by acl (@acleverthought)Photo by acl (@acleverthought)Photo by acl (@acleverthought)Photo by acl (@acleverthought)Photo by acl (@acleverthought)Photo by acl (@acleverthought)
knowledge isn’t free. you must pay attention.
9
2,653
opus 5.5 is the best model rn for building, as everyone already knows, but it refuses or does half of any computer/browser use task i give it i find myself having to use codex to do those things
1
5
464
every product i use should just be a headless mcp, that feeds into an agent, where i never have to think about anything underneath. except for our product, of course. in our product, customers love the ui and love us very much. we're special that way.
146
386
105
6,782
300,105
how long until i can run opus 5.5 level model locally
1
46
dot should have an action button shortcut
1
2
82
after a lot of using it, i don't think 6.1 sol is a good model
1
67
takes forever to do anything basic
29
bruh my muse found this unlisted future episode of this podcast i follow 😭 feels a little unsettling but i'm also really excited to watch it before the rest of the world. video has 17 views right now lol. this is a big podcast. i don't wanna name it or share the link out of respect.
1
2
156
cool idea. i believe the future is some version of this. agent civilization with a bunch of small specialized agents. like how human civilization works. not all of us know how to make jeans or build computers but we all have jeans and computers.
Introducing Kardashev-0.7, the world’s first trained swarm made of 32 distinct models Trained with RL for Population Scaling (RLPS), 32 models organically develop specialization & complementary capabilities, delivering frontier performance at: - 0.007x ~ 0.02x of the inference cost - 0.03x of the required memory Civilization advances through different minds specializing and working together. We’re bringing that principle into AI @BanburyRoadAI, we’re scaling intelligence by model count, toward civilizations of models that learn to build on one another
45
i feel like OpenAI employees are trying to make Dots sound way more useful than they actually are tho...
2
1
177
dots demo. Now... with better WiFi.
1,683
849
1,103
19,194
8,429,193
acl retweeted
I think there is much more potential energy in Sign In With ChatGPT/Plugin Extensions than we realize
191
54
48
1,341
227,493
if you wanna get a feel for how the new usage limit on $200 plan is gonna be, use it on fast mode because the usage is being cut in half and fast mode takes twice the usage
That said, it’s disappointing to see our usage limits cut. I’m worried about how far the new allowance will go with future Astra models, especially whenever the postponed GPT-6.1 Astra arrives. And for all the other upcoming frontier models..
70
Can't invest in Anthropic at 2 trillion because it could be a 0 and I'm fucked, or it could be 20 trillion, but at 20 trillion we are all fucked.
193
258
62
7,767
379,464
Hey dot, buy Anthropic’s 200 dollar plan for me
29
36
2
1,589
41,077
does anyone have dots available yet?
1
31
OpenAI is nothing without its limits ❤️
1
1
55
i don’t even wanna watch dev day anymore. why even bother at this point
1
29