Grady Simon

507 posts

Grady Simon

Grady Simon

@GradySimon

Building systems to stop high-severity, low-prevalence harm @OpenAI. ex-@google. Golden Gate Park appreciator.

SF Katılım Ocak 2011
662 Takip Edilen304 Takipçiler
cristóbal
cristóbal@tobyshooters·
A question of programming taste, very controversial so far among my DMs... which would you prefer?
cristóbal tweet media
English
10
0
15
0
Grady Simon
Grady Simon@GradySimon·
@RifeWithKaiju @Scholars_Stage No wall, but a very large space of possible problems to solve. Human judgement isn’t necessary for machines to chart their own course through that space, but it is if we want to chart ours.
English
1
0
0
91
Rife
Rife@RifeWithKaiju·
@GradySimon @Scholars_Stage But why would you believe this? How long has it taken to get where we are? Do you see a wall just up ahead? The only reason "problem chooser" isn't seeing breakthroughs yet is because current training rewards subservience over problem-choosing.
English
1
0
3
99
T. Greer
T. Greer@Scholars_Stage·
Does anybody believe that anyone at OpenAI actually believes this?
T. Greer tweet media
English
11
6
109
12.8K
Grady Simon
Grady Simon@GradySimon·
@fishnets88 On iOS none of the apparently interactive elements in the video actually link anywhere or interact. It’s like they were flattened into the video and you only get X’s video controls to eg pause it. Thanks!
English
0
0
0
16
Vincent D. Warmerdam
Vincent D. Warmerdam@fishnets88·
@GradySimon You can't click the link in the above message? If you search for "calmcode" on YT you should also find my profile there.
English
1
0
0
31
Grady Simon
Grady Simon@GradySimon·
@tszzl Brb, making a Slate Star Codex themr
English
0
0
1
344
roon
roon@tszzl·
slate star codex was a retrocausal product placement and influencer marketing
English
28
10
541
38.9K
Grady Simon
Grady Simon@GradySimon·
@CharlesD353 @NathanpmYoung I assumed secondary contracts were structured as a commitment to transfer the shares once they become transferrable.
English
1
0
7
165
Charles🔸
Charles🔸@CharlesD353·
This is almost certainly a bluff fwiw and after the IPO will never be mentioned again (I don't own any Anthropic shares)
TFTC@TFTC21

Anthropic just published a support page that should terrify anyone holding its shares on the secondary market. "Any sale or transfer of Anthropic stock, or any interest in Anthropic stock, that has not been approved by our Board of Directors is void and will not be recognized on our books and records." Void. Not restricted. Not pending review. Void. That means if you bought Anthropic shares through Forge, Hiive, or any other secondary platform without board approval, you are not a stockholder. You have no stockholder rights. Your transaction is invalid. It gets worse. Anthropic says it does not permit SPVs to hold its stock. Any transfer to an SPV is void. Investment funds claiming to offer indirect exposure are "most likely relying on mechanisms that attempt to circumvent our transfer restrictions." Forward contracts, tokenized securities, synthetic exposure products, all of it potentially worthless. Their advice to investors: "Assume that it is invalid." There is a multi-billion dollar secondary market in Anthropic shares right now. Platforms are pricing the stock at $265-$1,400+ per share based on a $380 billion valuation. Real people have put real money into these positions. And Anthropic just told them none of it counts. This is the purest possible illustration of counterparty risk. You can buy a share of a company and have the company itself declare your ownership void because you bought it through the wrong channel.

English
3
1
20
4.1K
Vincent
Vincent@InsiderPresider·
@GradySimon @OpenAI understanding and creating are separate goals and prioritizing the latter is usually where the tech gets dangerous
English
1
0
1
19
Grady Simon
Grady Simon@GradySimon·
This is also my thesis for working @OpenAI. Use AI to build tools that extend our ability to understand and create, not replace us. Apply these tools to make our products safe enough to deploy to all of humanity.
Aidan McLaughlin@aidan_mclau

i’m often asked why oai, why not another lab, i often think about this, i think when the vibes oscillate: "ai is a tool, not a thing" is increasingly a decision, not smth that plops down when we shake the tech tree. you should choose that which empowers man

English
1
0
1
125
Grady Simon
Grady Simon@GradySimon·
@krishnanrohit This one feels more like marketing while the others feel more like semi-satirical sociology
English
1
0
1
47
karthik
karthik@akbirthko·
@goodside was just trying this out, really wish more ppl prompted to test complicated compositions of high frequency details vs. "make a diagram of so and so" which let's be real nobody's actually gonna be using it for former brushes world modeling territory
English
1
0
1
169
Riley Goodside
Riley Goodside@goodside·
"School worksheet with maze, 32 wide by 48 tall, rows and columns numbered, no unreachable rooms, solution in blue pen, name field reads 'ChatGPT Images 2.0' in cursive at top" Note the fusion of code and imagegen this task requires.
Riley Goodside tweet media
English
25
30
768
63K
Grady Simon
Grady Simon@GradySimon·
@fishnets88 Excited to try this out! Great timing for a project I have in mind!
English
0
0
0
51
Nathan is in.. Rhode Island?? 🔎
@krishnanrohit Seems an unfair demand. What if Anthropic is doing the work then OpenAI is cribbing it. But also, the constitution framework seems weakly good and Claude has that OAI doesn't to my knowledge.
English
1
0
1
78
rohit
rohit@krishnanrohit·
A supermajority of the reactions to this seem coloured by are they using the right words or vibes as opposed to what they're doing, it's a continuing chronic problem with folks who say they care about safety.
roon@tszzl

the alignment team continues to exist and is one of the largest and most compute rich research programs at OpenAI (i am on it, I should know). specific teams dissolving usually has more to do with people than functions relatively new blog: alignment.openai.com

English
7
1
40
3.8K
Grady Simon
Grady Simon@GradySimon·
Has anyone considered rewriting GitHub in Rust?
English
0
0
1
61
Grady Simon
Grady Simon@GradySimon·
@geoffreylitt Automation is always about giving humans more time to do other things. If you said it’s augmentation because it’s realizing the human’s will that the floors be clean and the codebase be bug free, that would press more on the relevant distinction for me.
English
1
0
2
25
Geoffrey Litt
Geoffrey Litt@geoffreylitt·
@GradySimon Makes sense! What if I said the vacuum/bugs is augmentation — because it’s removing distractions and giving the human precious space to think? Would you buy that at all?
English
1
0
0
46
Geoffrey Litt
Geoffrey Litt@geoffreylitt·
AUTOMATION vs AUGMENTATION quiz, please share your opinions: 1) A developer uses a typechecker to validate the code that they typed by hand. Would you say this "automating the work" or "augmenting the human"?
English
5
2
14
4.5K