Ritchie Wong on X: "one recent learning habit is to ask perplexity (gpt image gen 2) to visualise long pieces for me and satya been cooking recently with x articles https://t.co/QTkVqfVBaL" / X
and satya been cooking recently with x articles
dad prefers kimi 2.5 for writing grading
helped dad to find a model to grade writing and he seems to like kimi the most
Ritchie Wong on X: "got to know richard sutton from andrej's podcast. the first post i saw on his feed already a banger then combining andrej's repeated use of the word 'entropy' in the podcast have a small new aha on what learning is: knowledge is like a swimming pool/ocean as accumulation of… https://t.co/qW0jZcfeg4" / X
then combining andrej's repeated use of the word 'entropy' in the podcast
have a small new aha on what learning is:
knowledge is like a swimming pool/ocean as accumulation of
Ritchie Wong on X: "i like how demis described the hybrid system in this all-in interview - something not just applicable in ai, but also in human learning it's either learning system or handcrafted system (experts) when we learn something new, we can start with a clearly defined predictable https://t.co/CNH93brbrD" / X
it's either learning system or handcrafted system (experts)
when we learn something new, we can start with a clearly defined predictable
Ritchie Wong on X: "i love this responsibility of anthropic head of devrel is to pioneer the definition of what world-class looks like i think that that's what all jobs have to do https://t.co/7B1fZObxBM" / X
i think that that's what all jobs have to do
Ritchie Wong on X: "love this convos on future of education esp, just-in-case learning vs just-in-time learning currently, most schools: learn XYZ first, maybe one day you will find situations to apply them what it should be: here're the problems/questions to solve, you don't know the answers https://t.co/Y6dQHId3wF" / X
love this convos on future of education
esp, just-in-case learning vs just-in-time learning
currently, most schools: learn XYZ first, maybe one day you will find situations to apply them
what it should be: here're the problems/questions to solve, you don't know the answers
Ritchie Wong on X: "just finished reading https://t.co/pSXwQS2XFa some quick thoughts 👇 https://t.co/8eLOwjHeAl" / X
just finished reading
https://t.co/pSXwQS2XFa
some quick thoughts 👇
Ritchie Wong on X: "one yardstick people talk about when deciding if ai has original creativity is whether it can come up with novel hypotheses to experiment and it's already happening 😲 while, as the team claims, it still requires time to do full validation with human trial, this is definitely" / X
one yardstick people talk about when deciding if ai has original creativity is whether it can come up with novel hypotheses to experiment
and it's already happening 😲
while, as the team claims, it still requires time to do full validation with human trial, this is definitely
Ritchie Wong on X: "one of the most alarming, depressing & awakening pieces about education & ai my favourite parts 👇" / X
one of the most alarming, depressing & awakening pieces about education & ai
my favourite parts 👇
Ritchie Wong on X: "the willingness to act without explicit validation, instruction, or even permissioneally the best time alive for people who seek no permission to make anything we want happen one of the most encouraging reads about agency, ai & permissionless" / X
the willingness to act without explicit validation, instruction, or even permissioneally the best time alive for people who seek no permission to make anything we want happen
one of the most encouraging reads about agency, ai & permissionless
Ritchie Wong on X: "just read anthoripic's report on values in ai seeing how these ai companies develop the value systems with the ai models reminds me a lot of how we humans develop ourselves or our children value https://t.co/7K5VXhzEuZ" / X
just read anthoripic's report on values in ai
seeing how these ai companies develop the value systems with the ai models reminds me a lot of how we humans develop ourselves or our children value
takeaways from anthropic's education report (2025)
just read anthropic's latest education report
while it's no secret that
- ai might potentially offload students' exercising & developing their cognitive skills
- CS students use claude far more than humanities, business students
- CS/maths students are engaged in