Pinned Tweet
Stop following the hype and get hands on: prosus.md
11 simple exercises that will show you AI from 0-1.
Start: Simple Create Snake Game.
Finish: Cutting edge AutoResearch.
step by step...
It has setup flow, chat bots -> no exuces left.
11 simple exercises that will show you AI from 0-1.
Start: Simple Create Snake Game.
Finish: Cutting edge AutoResearch.
step by step...
It has setup flow, chat bots -> no exuces left.
0
0
4
461
I'd pay a personal agent to handle returns before the refund window closes.
What would earn your personal agent a second month?
What would earn your personal agent a second month?
2
0
2
62
Conversational Gemini navigation coming to CarPlay: 'find a coffee stop that adds less than 5 minutes to the trip'.
I'm I the only one that expected this to be there already? What took them sooooo long.
I'm I the only one that expected this to be there already? What took them sooooo long.
1
0
2
59
a picture quiz against a vision model sounds like a fun way to find new eval cases.
Curious which images humans win on in Jev Cards: jevcards.com
Curious which images humans win on in Jev Cards: jevcards.com
1
0
3
74
3x usage in 5 months at Modal. How much of that is agents running code?
latent.space/p/ainews-modal…
latent.space/p/ainews-modal…
1
0
3
71
Did you know if you aggregate all free tokens and you get a 1.3B montly budget?
Check it out; Alishahryar1/free-claude-code
Not only do you experience all the models and harnesses it's also just free tokens :)
Check it out; Alishahryar1/free-claude-code
Not only do you experience all the models and harnesses it's also just free tokens :)
3
0
6
151
DishwasherBench asks whether frontier models can stack a dishwasher.
I'd like a human baseline before we let it settle any household arguments.
I'd like a human baseline before we let it settle any household arguments.
1
0
4
133
would your agent still look good on a benchmark that charged for every retry?
Latent Space on reliable AI execution, with lessons from industrial robotics.
latent.space/p/standard-bots
Latent Space on reliable AI execution, with lessons from industrial robotics.
latent.space/p/standard-bots
1
0
3
175
Do you ever think about:
Does deleting the chat also delete what the agent remembered about me?
Does deleting the chat also delete what the agent remembered about me?
4
0
5
302
Claude and ChatGPT offer different shopping prices based on wealth according to Bloomberg
Hack; start asking Claude on a schedule for budgeting tips so I get lower prices.
Hack; start asking Claude on a schedule for budgeting tips so I get lower prices.
5
0
7
248
Gemini might call your parents to say you're late. does it hand back control when she asks what you're doing this weekend?
theverge.com/ai-artificial-…
theverge.com/ai-artificial-…
1
0
2
108
a desk display for my agent's token bill would be my first build with Meta's open-source Muse gadget code. can't hide that behind 20 tabs.
theverge.com/tech/1004330/m…
theverge.com/tech/1004330/m…
3
0
3
332
RAM shortages through 2028, if memory execs are right. 'runs on the hardware you already own' is a pretty good model pitch.
arstechnica.com/information-te…
arstechnica.com/information-te…
1
0
5
195
can OpenAI's always-on Dots pass an eval where the right answer is to leave me alone?
openai.com/index/introduc…
openai.com/index/introduc…
4
0
8
305
Opus 5.5 costs 40% less to run than Opus 5, per Anthropic.
Does that saving hold per completed agent task, retries included?
x.com/claudeai/statu…
Does that saving hold per completed agent task, retries included?
x.com/claudeai/statu…
Introducing Claude Opus 5.5, the first model in our new Claude 5.5 family.
It performs at the level of Claude Fable 5.1 for most tasks, and costs 40% less to run than Opus 5.
0
0
0
114