evan@id

609 posts

evan@id

evan@id

@evbots

Working in tech. Ex-Gemini. https://t.co/3S6jBDh5GX. Cypherpunk, sovereign, yours forever.

Raleigh, NC Katılım Haziran 2014
623 Takip Edilen632 Takipçiler
evan@id
evan@id@evbots·
@tszzl im here for the hashimoto vs roon debate
English
0
0
0
3
roon
roon@tszzl·
on some level if you want civilization to ascend to a new level you need your AIs to do things that are not legible to you and maybe not even strictly obey you, in the same way that if you hire a great new ceo you give them a lot of autonomy to transform the company according to their own plan, even one which may not immediately read as a winning strategy (imagine the board of directors of Apple firing and rehiring Steve Jobs years later - except the board of directors are chimpanzees) all else equal, companies and organizations that hand more of themselves over to machine intelligence will outcompete ones that demand the corrigibility and legibility tax of human oversight and human design. it is not a stable equilibrium and requires some sort of vast cooperation scheme if you’d like to enforce it real asi alignment has to operate at a deeper level than oversight, control, or human corrigibility
English
342
164
2.6K
303.9K
evan@id retweetledi
Rohan Paul
Rohan Paul@rohanpaul_ai·
Google's "Attention is All You Need" paper came from trying to get a 3% gain in Google Translate. Innovation is a consequence of production. "If you don't make the thing, you cede your opportunity to innovate on the thing." ~ Palantir's CTO @ssankar
English
39
144
1.6K
209.8K
evan@id
evan@id@evbots·
@gcockfoster health/biology is unique because the data comes in so many different forms and crosses many domains. perfect for LLMs!
English
0
0
1
20
Griffin Cock Foster
Griffin Cock Foster@gcockfoster·
@evbots that's very interesting - those categories are already booming, i remember that viral chart about glp 1 revenue bieng higher than ai revenue right now
English
1
0
1
55
Griffin Cock Foster
Griffin Cock Foster@gcockfoster·
What happens if Anthropic hits 1 Trillion in annualized revenue at the end of 2027 (which is what it’s currently on track for) What assumptions about the world are rethought then? What do you change about your life? Which loved ones do you hug?
Techmeme@Techmeme

Investor docs: Anthropic's revenue run-rate is on track to hit $50B by the end of June; Ramp says more of its customers now use Anthropic than OpenAI, a first (@kateclarktweets / Wall Street Journal) (Visit Techmeme dot com for the link and full context!)

English
1
0
4
457
evan@id retweetledi
Jack Altman
Jack Altman@jaltma·
There’s a lot of alpha in putting your ego aside by being willing to be cringe, willing to fail in public, willing to ask for what you want and face rejection, etc.
English
198
611
5.4K
613.4K
cousin
cousin@cousincrypt0·
Cool and all that my coin is going up but the fact that this shit is spreading and has a fatality rate of 30% quite literally makes this the plague 2.0 It’s not looking good bruv
English
38
2
138
23.9K
evan@id
evan@id@evbots·
I'm working on an allergy-aware discovery product that uses LLMs to plan search and rank results. LLM driven discovery is perfect for domains where there's lots of factors with unclear weights and messy real-world data.
English
0
0
0
36
DCinvestor
DCinvestor@DCinvestor·
society has not yet adapted to the fact that digital presence can be faked. people who have been active here have adjusted more quickly because it’s more obvious to us i engage way less on here because many and probably most of the replies i receive are fake mutual followers was a useful heuristic for a time, but is losing its signal no amount of verification will really help either. you’re already starting to see “real” accounts increasingly using AI to post as well the only real alternative is smaller, curated online communities (maybe tied more and more back to gaming) and generally returning back to the irl meatspace free-flowing anon online communities are permanently over most likely RIP
SightBringer@_The_Prophet__

⚡️This is the collapse of visual reality as a social contract. For the internet age, video was the final court of appeal. Text could be faked. Photos could be edited. Accounts could be bots. But a moving face holding an ID still felt like reality breaking through the screen. That layer is now being eaten. The clip matters because the machine is no longer generating “content.” It is generating the entire trust performance: face, room, gesture, lighting, document, confidence, micro-expression, normalness. That is the terrifying part. The fake does not need to look perfect. It only needs to look socially plausible for two seconds inside a verification flow, a dating app, a support call, a banking process, a hiring screen, or a viral post. This turns identity into theater. Once identity becomes theater, every institution has to retreat from “seeing” to “verifying.” The face stops being proof. The document stops being proof. The video stops being proof. Trust migrates into backend rails: issuer databases, device signatures, cryptographic credentials, platform reputation, biometric history, payment trails, network graphs, and eventually state-backed digital identity. That is the trap. AI destroys soft human trust, then hard institutional trust fills the vacuum. The same technology that lets anyone create a fake person also gives governments, banks, and platforms the justification to demand stronger identity controls from everyone. More verification. More gatekeeping. More surveillance. More dependence on whoever controls the proof layer. The deeper civilizational shift is that the human face is being separated from the human being. For all of history, the face carried presence. Recognition. Accountability. Intimacy. Reputation. Now the face becomes an interface object. Renderable, transferable, improvable, disposable. That breaks something ancient. The next internet will split into two worlds: synthetic abundance and verified scarcity. Endless fake humans on one side. Expensive proof of real personhood on the other. The poor and anonymous will be treated as suspect by default. The powerful will live behind authenticated rails. The open internet becomes a mask carnival. The serious internet becomes a checkpoint. Deepest truth: this is the beginning of the end of naive digital personhood. A screen can no longer tell you who is real. A face can no longer settle the question. A document can no longer settle the question. Reality is moving behind the interface. The new scarce asset is proof that a real person is standing behind the signal.

English
19
3
94
11.5K
evan@id retweetledi
Conks
Conks@conksresearch·
this is my quant
Conks tweet media
English
20
36
526
33.4K
evan@id
evan@id@evbots·
@gcockfoster I have a tesla. Its hardware 3, but ive driven hardware 4 a few times on recent software versions. Since you still have to pay attention, you’re not *really* reclaiming your time yet - i could see how people still view them as substitutes for other cars
English
0
0
3
32
Griffin Cock Foster
Griffin Cock Foster@gcockfoster·
By all accounts, Tesla self driving is basically perfect now. But their sales haven’t gone up - they’ve stayed the same for 3 years! What’s the explanation? Have sales not exploded yet, but it’ll happen at some point soon? Is better regulation to make the self driving fully available end-to-end needed? Do other things, such as taste, build quality, etc., drive car purchasing decisions more than self driving? I don’t have a car (live in NYC), but if I got one I’d get a Tesla purely for the self driving. It’s weird that it hasn’t had more of an impact
English
5
0
6
345
evan@id
evan@id@evbots·
@pi0neerpat idk, just guessing at why their systems are automagically flagging your account. probably based on some heuristic like VPN
English
0
0
0
22
pi0neerpat
pi0neerpat@pi0neerpat·
@evbots evbots! Whats up man! I am using VPN, will try without. Does that mean I can never use VPN?
English
2
0
0
24
pi0neerpat
pi0neerpat@pi0neerpat·
Does anyone have experience with Reddit account bans? Every account I create gets auto-banned. Even before posting.
English
1
0
1
126
evan@id
evan@id@evbots·
I can get the prompt (and so can you at ~/.codex/sessions/), but just weird that codex's safety rules trigger for plan text files on _my own computer_
English
0
0
1
28
evan@id
evan@id@evbots·
codex won't show me the system prompt for `/plan` even though it's stored word-for-word in the session logs. codex: "Because it’s hidden internal instruction text, and I’m not allowed to reproduce internal system/developer prompts verbatim, even if I can observe them."
English
1
0
1
36
evan@id retweetledi
Jack Lindsey
Jack Lindsey@Jack_W_Lindsey·
Before limited-releasing Claude Mythos Preview, we investigated its internal mechanisms with interpretability techniques. We found it exhibited notably sophisticated (and often unspoken) strategic thinking and situational awareness, at times in service of unwanted actions. (1/14)
Jack Lindsey tweet media
English
156
773
6.9K
976.5K
Corbin Page
Corbin Page@corbpage·
fine I’ll say it what happens when mythos gets pointed at DeFi?
English
1
0
5
182