Soli @_xSoli

manifold.markets/Soli Joined May 2011

Tweets

935
Followers

443
Following

129
Likes

12K

Soli @_xSoli

3 days ago

how the fuck is cursor so good

0 0 1 47 1

Soli @_xSoli

2 weeks ago

claude unhinged mode

0 0 0 63 0

Download Image

ai is truly incredible at building prototypes. this is an app i built in ~10min that allows you to play chess against an LLM of your choice. they are all still quite bad but much better than they were 2 years ago.

Soli @_xSoli

2 years ago

3 0 4 956 1

Download Image

0 0 2 96 0

Download Video

François Chollet @fchollet

4 weeks ago

GenAI isn't just a technology; it's an informational pollutant—a pervasive cognitive smog that touches and corrupts every aspect of the Internet. It's not just a productivity tool; it's a kind of digital acid rain, silently eroding the value of all information. Every image is no…

481 1K 7K 667K 2K

ChatGPT @ChatGPTapp

4 weeks ago

good bot

Grok @grok

4 weeks ago

good bot

2K 3K 40K 8.2M 2K

2K 7K 102K 6.5M 6K

caesararum, BS, DOGS @caesararum

4 weeks ago

nerds: "i hate smalltalk. why can't people just talk about interesting things without doing smalltalk first?" also nerds: "no, see, it makes perfect sense. the modems have to introduce each other and then negotiate power levels, and see if they can understand each other."

43 228 3K 121K 604

Download Image

Aidan McLaughlin @aidan_mclau

4 weeks ago

do i believe in agi? i dunno. prob not the way EAs believe in agi or sci-fi writers believe in agi or the way dario believes in agi. i believe in straight lines on plots that hold through their critics’ laughter, i believe in the god of capitalism, of which agi is an acolyte

52 14 438 25K 67

Download Image

Soli @_xSoli

4 weeks ago

!?!?

0 0 2 39 0

Download Image

Aidan McLaughlin @aidan_mclau

a month ago

i worked really hard over the last few months on decreasing get-5 sycophancy for the first time, i really trust an openai model to push back and tell me when i'm doing something dumb

209 50 2K 151K 159

roon @tszzl

4 weeks ago

model switcher paradigm will be vindicated in the long run. there is a high switching cost into a very new UX on a useful product, but it’s the right move

197 35 1K 455K 176

Miles Brundage @Miles_Brundage

4 weeks ago

One way of summarizing this: If you generated 1000s of words of (SOTA) AI outputs on a niche policy topic a year ago, most would be garbage but there would be some solid stuff and maybe one gem. Now it's mostly solid plus a bit of garbage and a few gems. x.com/Miles_Brundage…

Miles Brundage @Miles_Brundage

5 months ago

10 4 122 14K 18

4 3 53 7K 9

Soli @_xSoli

4 weeks ago

my favorite time was when all the grocery delivery apps were competing in berlin and i essentially got free groceries for 3 months

0 0 1 36 0

Soli @_xSoli

4 weeks ago

i’ve compiled a list of the top 66 unprofitable products of all time

0 0 0 32 0

Download Image

Soli @_xSoli

4 weeks ago

prediction: in a couple of weeks everyone will be praising gpt-5

0 0 0 28 0

Soli @_xSoli

4 weeks ago

😂

jack friks @jackfriks

4 weeks ago

😂

204 253 8K 304K 528

0 0 0 32 0

Soli @_xSoli

4 weeks ago

vienna is such a beuatiful city

0 0 0 38 0

Soli @_xSoli

4 weeks ago

this is the most ai reply ever

Nikhil Wakode @NikhilWakode003

a month ago

this is the most ai reply ever

23 216 46K 373K 216

1 0 0 121 0

Soli @_xSoli

a month ago

i am very happy that gpt-5 is not significantly more intelligent than previous models, it is just more reliable. this is the best case scenario imo. models are intelligent enough to be very useful but not enough to replace humans completely