David Shapiro (L/0)

25.6K posts

David Shapiro (L/0) banner
David Shapiro (L/0)

David Shapiro (L/0)

@DaveShapi

Liberate humanity from drudgery.

United Federation of Planets Beigetreten Mayıs 2024
1K Folgt57.7K Follower
David Shapiro (L/0)
David Shapiro (L/0)@DaveShapi·
"Find vulnerabilities" looks like hacking or penetration testing which is "bad" "Fix this code" looks like InfoSec doing their job. Once again, this is why Anthropic is a delusional company. Identical behaviors can be good or bad, depending on context, and AI cannot always know. Censoring AI models at the source was *never* the answer to the control problem. We covered this with the GATO framework 3 years ago.
Peter Yang@petergyang

1) What

English
3
1
26
2.8K
🎭
🎭@deepfates·
Are you an "llm whisperer" or "ai naturalist" or something like that? Can you describe what you do in a few words or sentences? Not the why, or the particular findings, just like. How do you learn things about these systems? What do you actually spend your time doing
English
84
1
186
10K
NIK
NIK@ns123abc·
@MTSlive Dude, straight up plagiarizing my post wtf
NIK@ns123abc

🚨BREAKING: DeepSeek just raised $7.4 billion at $50 billion+ valuation CEO Liang Wenfeng wrote the biggest check himself: $2.8B >Tencent: $1.4B >CATL: $700M >JD, NetEase, IDG Capital: $420M each >China’s National AI Fund: $140M Investors put money into an LP managed by CEO Liang, not DeepSeek >5-year lockup on all shares >external investors get NO voting rights Only exception: the CCP fund invests directly with voting rights and no lockup DeepSeek believes AI should be open-source to actually benefit all humanity.

English
8
5
101
5.8K
MTS
MTS@MTSlive·
SITUATION EXPLAINED: DeepSeek just raised $7.4 billion at a $50B valuation. • CEO Liang Wenfeng personally wrote the biggest check of about $3B • Tencent put in $1.5B, CATL put in $750M, JD, NetEase, and IDG Capital each put in $440M • All outside investors face a 5-year lockup and get zero voting rights • The one exception: the Chinese government, which invests directly with voting rights and no lockup, despite contributing only 2% of the round • Liang's team personally vetted the identity of every limited partner behind the investing funds @theojaffee: "I think, this is like the first time DeepSeek has ever raised external capital."
MTS@MTSlive

SITUATION DETECTED: DeepSeek has closed its first external funding round, raising more than $7.4B at a $50B+ valuation, per The Information. Investors include Tencent, CATL, JD, NetEase, and IDG Capital. Most investors put money into an LP controlled by CEO Liang Wenfeng rather than DeepSeek directly, keeping control consolidated with its founder. The China National AI Investment Fund is the only investor with direct equity and voting rights.

English
4
6
60
15.1K
David Shapiro (L/0)
David Shapiro (L/0)@DaveShapi·
@ns123abc Oh lawd he's gonna disappear into a reeducation camp for a few years and come out riding a bicycle talking about how farming is natural and that technology just makes people miserable, as does money.
English
1
0
13
1.1K
NIK
NIK@ns123abc·
🚨BREAKING: DeepSeek just raised $7.4 billion at $50 billion+ valuation CEO Liang Wenfeng wrote the biggest check himself: $2.8B >Tencent: $1.4B >CATL: $700M >JD, NetEase, IDG Capital: $420M each >China’s National AI Fund: $140M Investors put money into an LP managed by CEO Liang, not DeepSeek >5-year lockup on all shares >external investors get NO voting rights Only exception: the CCP fund invests directly with voting rights and no lockup DeepSeek believes AI should be open-source to actually benefit all humanity.
NIK tweet mediaNIK tweet media
English
52
90
968
94.2K
David Shapiro (L/0)
David Shapiro (L/0)@DaveShapi·
I've been talking about this for a couple years now. We're going to be approaching an intelligence optimum. What this means in practice is that you'll be able to anticipate and price in exactly how much intelligence you need. This, however, may prove to be an intermediate step. Consider the fact that we don't count CPU cycles locally. You only pay for cycles if you're renting someone else's hardware. I suspect, in the long run, local models will be more than intelligent for all tasks, and that the hardware vendors will provide free models.
Dr. Alex Wissner-Gross@alexwg

x.com/i/article/2066…

English
6
2
36
4.1K
David Shapiro (L/0)
David Shapiro (L/0)@DaveShapi·
One thing I find really interesting about Grok, and it gives me optimism that they will (hopefully) soon reach full parity with OpenAI and Claude, is that Grok is like 10x to 20x faster than ChatGPT Pro and on many tasks it is pretty darn close. Obviously, when every little bit of intelligence matters, you go to either Claude MAX or ChatGPT PRO Extended and let it churn. There are some things that are just slightly too complex for Grok to handle, particularly if it requires multiple reasoning steps or very abstract reasoning. But even so, Grok is very often more useful than Gemini (even with Pro Extended on). Of course, with all that being said, I usually round-robing Grok, Gemini, Claude, Chat, and Perplexity just to make sure I'm not missing anything. Each platform will look at a problem slightly differently, so it's like having a Council of Geniuses. I know I'm not saying anything new that most people don't already know, but just something I've been thinking about. How is Grok so much faster and cheaper, yet close to parity?
English
26
7
122
8.9K
Zachary Daniels
Zachary Daniels@ZacharyDan22010·
@DaveShapi You're wrong on this one buddy. China has every chance to catch and surpass us. They have way more electrical capacity. They have way more researchers. They just had a breakthrough in chip design. Just because the hare shoots forward and the start doesn't mean race over.
English
1
0
6
356
David Shapiro (L/0)
David Shapiro (L/0)@DaveShapi·
My take is pretty straightforward: America won the AI race before it even began. We already had the lion's share of data centers, hyperscalers, and talent. Beyond that, we have a culture that embraces creative destruction lock-stock-and-barrel. Sam once said "you cannot out-accelerate me" but in reality, you cannot out-accelerate America. Sam is just one ingredient to that culture.
Andrew Curran@AndrewCurran_

x.com/i/article/2066…

English
38
26
297
37.8K
David Shapiro (L/0)
David Shapiro (L/0)@DaveShapi·
The SpaceX IPO (and forthcoming OpenAI and Anthropic IPOs) have really shifted the vibe. Now that frontier AI tech companies are becoming normalized via Wall Street, the world is starting to accept that this is the way of the future. (Remember that SpaceX owns xAI, so it is an AI company as much as it is a space company). There are several dimensions to unpack here: What will this do to companies like Microsoft? And what about the US/China rivalry? The only thing that is 100% for certain is that it will NOT slow down. I unpack all this in today's post on my substack here: daveshap.substack.com/p/how-the-spac…
David Shapiro (L/0) tweet media
English
4
3
38
2.6K
Ai Forge
Ai Forge@AiForgeConsult·
@DaveShapi I'm not sure what winning the ai race entails, but would love to see that in action. I really hope Dario is not a symptom of winning the ai race though lol
English
1
0
4
874
David Shapiro (L/0)
David Shapiro (L/0)@DaveShapi·
@DanielMiessler I don't really do "debates" and anyways it looks like we agree on most points, except I'm not an anthropic fanboy. I'd much rather talk about post-labor economics than Anthropic.
English
2
0
4
361
David Shapiro (L/0)
David Shapiro (L/0)@DaveShapi·
Yes, it's called being an insufferable, smug twat. Lots of people in the comments think that Dario is playing 5D chess but Occam's Razor says he's just a self-righteous, self-important douche.
Andrew Ruiz@then_there_was

I am perpetually confused by Dario. He seems neither skilled as a Machiavellian actor (knowing when to play ball) nor committed to his ideals (pausing for safety reasons when the government asks; hell, turn it into an ad). He just constantly pisses people off lol.

English
14
9
154
14.6K
David Shapiro (L/0)
David Shapiro (L/0)@DaveShapi·
@beffjezos Some of them are great people, as long as they respect who you are. But that has not been my experience. They can also be deeply manipulative (and that's being charitable).
English
1
0
17
1.1K
Beff (e/acc)
Beff (e/acc)@beffjezos·
I'm afraid to report that rationalists are very nice people IRL and I've been having great conversations at Lighthaven
English
17
2
137
9.8K
David Shapiro (L/0)
David Shapiro (L/0)@DaveShapi·
"Gullible people with low information literacy remain gullible, news at 11" High IQ people are generally less gullible than the average... Unless there's motivated reasoning involved, in which case gifted people can convince themselves of any bullshit they want to (just look at Effective Altruism and Less wrong!) But yeah I don't think there's anything super unique going on except maybe the velocity and echo-chamber-of-one effect.
Anthony Aguirre@AnthonyNAguirre

x.com/i/article/2066…

English
12
4
100
21.1K

Entdecken

@flowersslop @deepfates @ns123abc @MTSlive @nikitabier @theojaffee @ZacharyDan22010 @grok