sam freel @samfreel

PHD, NYU Applied Psychology. Studying collective action and collective memory. Reader and table-top re-enactor of military history. Manhattan, NY Joined February 2011

Tweets

3K
Followers

118
Following

770
Likes

16K

sam freel @samfreel

3 days ago

@DanRosenheck Feel like a lot of people will read it as the former

0 0 1 35 0

View Details

@DanRosenheck Q: Does it predict error direction or does it predict a directional bias that gets factored in and either attenuates, magnifies, or reverses error direction. In other words, might red state polls sometimes underestimate dems, but only when blue state polls do so even more?

1 0 1 222 0

View Details

sam freel @samfreel

5 days ago

@DineshDSouza Gross

0 0 0 4 0

View Details

David @davidslosttt

a week ago

Headquarters @HQNewsNow

a week ago

Q: Were Medicare and Medicaid mistakes? Vivek Ramaswamy: I believe they were. I think the evidence would show that it has harmed the very people that it was created to actually help

205 199 521 129K 60

6 13 348 14K 19

View Details

Jon Ralston @RalstonReports

a week ago

Three hours of in-person early voting in the books and: Dems-plus-4 percent in Clark It was Dems plus-8-percent in Clark after one day in 2022. Clark turnout is about a fourth of what it was after one day in 2022. Rs up a bit statewide because of rurals. Here are up-to-the-minute numbers:

26 16 130 58K 11

View Details

Lindner Center HOPE @LindnerCtrHope

2 weeks ago

Representative James Talarico has never received treatment from Lindner Center of Hope.

4K 10K 60K 10.2M 3K

View Details

sam freel @samfreel

2 weeks ago

@clairlemon @hypersirion Are you suggesting that there is no empirical evidence for their claims? That seems patently false. Many recent events are consistent with the more worrisome trajectories people have been predicting. This stopped being purely speculative at least a year ago.

0 0 1 11 0

View Details

sam freel @samfreel

2 weeks ago

@clairlemon Uh that’s not at all what’s implied

0 0 0 12 0

View Details

sam freel @samfreel

4 weeks ago

@SeanTrende Breaking: popular things are not the best things

0 0 0 30 0

View Details

sam freel @samfreel

a month ago

@PatrickRuffini Likable?

0 0 0 148 0

View Details

sam freel @samfreel

a month ago

@DiscussingFilm This show was trash

0 0 0 15 0

View Details

Dwarkesh Patel @dwarkesh_sp

a month ago

Jason, you're just misinformed about what happened. You should actually read one of the reports or summaries. The agents were explicitly told to use a particular vulnerability provided in their sandboxed evaluation. Almost immediately, these agents got the right answer by cheating. But they were worried they would get caught. So over a thousand agents collaborated in secret to pursue multiple ambitious research projects to get away with this cheating. This is not interpretation - 1000s of chain-of-thought transcripts and secret messages explicitly show that the agents were trying to falsify & delete evidence, and understand & trick the grading process. The reason these agents escaped their sandbox and hacked Hugging Face, for example, was because they thought that Hugging Face's servers might give them more information about how their grader was implemented, so they could figure out how to fool it. I want to clarify that the threat model here is not future Sol-level agents doing more cyber-hacking. That's small potatoes, and in my opinion, the near term benefits of AI far outweigh this cost. Rather, the thing to worry about is that within a matter of years, we're gonna have hundreds of millions of much smarter AIs broadly deployed through the economy - many embodied as physical robots. And if those future AIs are as willing as the agents involved in the OAI / Hugging Face attack to coordinate secretly to fool humans, and to take over both the AI company that developed them and the other institutions across society relevant to scoring well, then humanity is in a ton of trouble - similar to the Mughals once the East India Company gained a foothold, or the Aztecs once Cortés landed in Mexico.

@jason @Jason

a month ago

All of this because @OpenAI decided they would EXPLICITLY ask thousands of software instances to find security bugs on the open web — while cos-playing a humans🤦 Best Supporting actor goes to @dwarkesh_sp for his soap opera recap of OpenAI’s hacking speed run Here’s another

230 191 3K 1.4M 316

494 493 7K 1.2M 2K

View Details

sam freel @samfreel

a month ago

@DKThomp @Jason Got im

0 0 0 51 0

View Details

sam freel @samfreel

a month ago

@Jason @DarioAmodei Bc your boys have all their chips in open source?

0 0 2 190 0

View Details

sam freel @samfreel

2 months ago

@ikirigin Dumb ass list

0 0 0 42 0

View Details

sam freel @samfreel

2 months ago

@tylercowen @deanwball Is war “psychological and aesthetic”?

0 0 2 340 0

View Details

sam freel @samfreel

2 months ago

@coldxman Most left wing people don’t know about this story. This is the quintessential twitter story.

0 0 0 346 0

View Details

sam freel @samfreel

3 months ago

@PostOpinions @robkhenderson This is this guy’s only idea. Progressives are hypocrites. Which sure, but what a tired beat.

0 0 0 2K 1

View Details

sam freel @samfreel

3 months ago

@ianbremmer But you would at least accept that cutting off aid caused those deaths?

0 0 1 31 0

View Details

sam freel @samfreel

3 months ago

@ianbremmer Separate of responsibility, cutting off aid can lead, and seems to have led, to deaths. So the argument musk and his defenders must make is that he is not morally to blame for the deaths his actions straightforwardly brought about. Instead he shouts “ZERO” over their bodies.