Zhengdong
@zhengdongwang
Followers
2K
Following
286
Media
9
Statuses
95
natural history @GoogleDeepMind / consumer.
London, England
Joined May 2016
Spanish colonists establish a fort in the American Southwest, in a place that they call “Tucson.”
13
72
1K
it was so fun to collaborate with @sebkrier and @IFP writing about evaluations, read all the awesome proposals below!
THE LAUNCH SEQUENCE A new collection of 16 essays on how to accelerate AI for science & security: https://t.co/pKKwgRSdS7 The AI revolution is already delivering enormous consumer benefits. But AI progress won't automatically solve humanity's most important problems first. To
0
3
24
in which i get to play my version of 'overrated or underrated' with tyler cowen listen to tyler's full talk below
In a discussion on AI and economics at DeepMind, @tylercowen argued that while model progress is fast, real-world frictions mean the "slow takeoff" for broad living standards is a tough, possibly decades-long slog. Great questions by @zhengdongwang too!
10
10
149
this is a must read, stories from recent india you won't find anywhere else
NEW ESSAY: "Dispatches from India" A year ago I was living in New Delhi, India 🇮🇳. I traveled across 10 states and 30 cities+villages writing for The Economist, including a cover story on the economy Since then I’ve been chewing on how to capture what I learned It still
0
0
6
I wrote some fiction in the style of AI 2027. It combines the parts of AI 2027 and AI as Normal Technology that resonate with me most. Come for the predictions, stay for the animal parables!
1
7
46
Today is APOD's 30th birthday! In celebration, today's picture uses past APODs as tiles arranged to create a single pixelated image that might remind you of one of the most well-known and evocative depictions of planet Earth's night sky. In fact, this Starry Night consists of
17
130
545
Watching the end of Andor, then Rogue One, then A New Hope rocks, because the plot and the aesthetic are basically seamless, but everyone in the galaxy gets way stupider over the course of maybe a week
188
3K
79K
math pope: "perplexity"
0
0
7
conflict of interest obviously; i haven't been shilling enough and quiet period is over
0
0
61
guys buy google stock, it's such a good deal--- we've got waymo, we've got youtube, we've got the entire model pareto front, and we've even got 14% of anthropic and exposure to ssi, if that's what you care about
34
9
784
Andor dares to ask, “What if Star Wars was good?”
23
541
6K
this empirical study from @saffronhuang is so cool--- we desperately need more research like it!
Really proud and excited to release this work on empirically measuring AI values “in the wild” — understanding, analyzing and taxonomizing what values guide model outputs in real interactions with real users. There is a lot of work on training models to follow particular
0
0
2
Many years later, as he faced the firing squad, Colonel Aureliano Buendía was to remember that distant afternoon when his father took him to Europa to enlist as an ice harvester.
11
193
2K
Introducing MafiaBench: an LLM eval testing models' abilities to persuade, deceive and engage in strategic play using the social deduction game of Mafia I ran a 450 game Swiss tournament tldr: 4o blows the other models away. Claudes struggle Link and more details below 👇
4
4
33
@yuanyi_z Yale’s loss is somehow also Toronto’s loss.
3
2
53
two of the most interesting writers today giving you pure consumer surplus!
new podcast up on BUILDERS VS. PRIGS ft. @arjun_ramani3, economist and my favorite fellow niche-tech-ideology-watcher I always learn a ton from Arjun. here's what we covered: 00:49 Zuck's new look 02:05 PG's new essay on wokeness 07:08 origins of the "tech right" 16:02 why tech
1
0
7