kelmgren Profile Banner
Karson Elmgren Profile
Karson Elmgren

@kelmgren

Followers
1K
Following
4K
Statuses
2K

徐森凱 | now AI policy research @RANDCorporation prev @CSETGeorgetown @OpenAI &c | 🗣 中文 русский … | 横眉冷对千夫点赞

Old Gold Mtn.
Joined March 2011
Don't wanna be here? Send us removal request.
@kelmgren
Karson Elmgren
4 days
check it out:
0
0
2
@kelmgren
Karson Elmgren
9 days
@andrew_n_carr congrats!
0
0
2
@kelmgren
Karson Elmgren
12 days
RT @ohlennart: Here, now you hear it from a US company itself: it turns out they also had compute efficiency improvements but did not tout…
0
99
0
@kelmgren
Karson Elmgren
12 days
RT @mattsheehan88: Very happy to share this conversation w/ my colleague @SamWinterLevy about DeepSeek and the implications for 🇨🇳-🇺🇸. (lin…
0
8
0
@kelmgren
Karson Elmgren
13 days
0
0
1
@kelmgren
Karson Elmgren
14 days
see, if you're gonna give us your unhinged DeepSeek theories, at least make them entertaining
@soupS0S
soup
14 days
deepseek's revenue model is not to make money through ai models but to short relevant ai / chip firms with its hedge fund arm that makes the big bucks whenever it releases a new open license model and pump that money back into the research team
0
0
8
@kelmgren
Karson Elmgren
14 days
@mkwitzke oof, this was so relatable...
0
0
1
@kelmgren
Karson Elmgren
17 days
RT @fiiiiiist: Some reflections on DeepSeek’s impressive success and AI scaling. General trends across models tell us that algorithmic im…
0
23
0
@kelmgren
Karson Elmgren
20 days
RT @Saad97Siddiqui: The US & China may be strategic competitors in AI, but neither wants to see dangerous AI systems that boost rogue actor…
0
18
0
@kelmgren
Karson Elmgren
21 days
RT @_achan96_: Whether 2025/6/7/8/etc is the year of AI agents, we will need tools to help us unlock their benefits and manage their risks.…
0
32
0
@kelmgren
Karson Elmgren
24 days
@mattsheehan88 you're so brave for this
1
0
2
@kelmgren
Karson Elmgren
24 days
RT @mattsheehan88: Four years ago I gathered data on how often the People's Daily (top CCP newspaper) mentioned different emerging technolo…
0
2
0
@kelmgren
Karson Elmgren
1 month
the greatest contribution of this project might actually be the invention of this button tho
Tweet media one
@natfriedman
Nat Friedman
1 month
We did it! We tested 300 Bay Area foods for plastic chemicals. We found some interesting surprises. Top 5 findings in our test results: 1. Our tests found plastic chemicals in 86% of all foods, with phthalates in 73% of the tested products and bisphenols in 22%. It's everywhere. 2. We detected phthalates in most baby foods and prenatal vitamins. 3. Hot foods which spend 45 minutes in takeout containers have 34% higher levels of plastic chemicals than the same dishes tested directly from the restaurant. 4. The 1950s Army rations we tested contained surprisingly high levels of plastic chemicals. 5. Almost every single one of the foods we tested are within both US FDA and EU EFSA regulations. Check out our full results below.
0
0
7
@kelmgren
Karson Elmgren
1 month
tbh if I'd let him he might've just sat with me until we had one working, from scratch. really admirable when brilliant ppl are so modest and generous with their time he's lowkey kinda funny too. just a chill dude
0
1
58
@kelmgren
Karson Elmgren
2 months
RT @karpathy: DeepSeek (Chinese AI co) making it look easy today with an open weights release of a frontier-grade LLM trained on a joke of…
0
3K
0
@kelmgren
Karson Elmgren
2 months
@Liface 哈哈哈 what a throwback - the good old days!
0
0
0
@kelmgren
Karson Elmgren
2 months
@hustlerone4 it would be very convenient if the One True Morality were both obvious and easy to fully specify! unfortunately I think it is neither, if it even exists
0
0
1
@kelmgren
Karson Elmgren
2 months
@chaeronanaut yep, as long as the AI is epistemically humble about its current understanding of "the good" (which I don't think is the case for current systems or will be in the future by default) - this is Stuart Russell's basic plan for alignment, see eg
0
0
1