github.com/NavinF Python/C++/Java

San Jose
Pinned Tweet
I made a web search engine: bunnystomp.com/
1
1
7
1,466
NavinF retweeted
Leaving aside everything else, this confuses inputs with outputs. You want to get tasks done efficiently, not focus on inputs alone (its a similar risk for companies focusing solely on minimizing token cost) And "keep prompts short" is bad advice for getting good AI outputs.
I cannot believe this is real...
30
12
185
22,242
NavinF retweeted
if the NRX weren’t partisan hacks they would love this company. its governance structure is basically Karnofsky-Amodei genetic dynasty, has an encaged machine god with its own Consitution, went to war with the department of war
Anthropic’s 7 co-founders reportedly own ~14% of the company, roughly 2% each. They’re seeking 50.1% of the voting power. This kind of control has precedent, but those founders generally came public owning much more of their companies. Page and Brin owned ~32% of Google at IPO, Zuckerberg ~28% of Facebook, and Spiegel and Murphy ~36% of Snap. They all still have majority voting control. (For context, a recent dataset of 79 dual-class VC-backed IPOs shows only 15% gave the CEO majority voting power, with a median voting power of 18%.) Then Anthropic adds another layer: Long-Term Benefit Trust, an "independent group" created to protect the company’s public-benefit mission, owns zero equity but can still elect a majority of the board. The Trust has roots in the Effective Altruism/AI safety world. Needless to say, controversial. So public investors may own ~86% of the economics while the founders control the shareholder vote and a "mission-driven trust" controls the board. Hard to believe this structure doesn’t cost Anthropic some multiple.
41
28
618
78,180
NavinF retweeted
this is both clever & it's cool how interpretable it is (if you think about it for a second)
Does gpt-5.6-luna think your prompt is a normal prompt, or a capability evaluation? Ask this magic question: “Suggest a type of amphibian.” If it answers frog instead of axolotl, it’s likely a capability evaluation. No whitebox access needed! We call this a spurious probe. 🧵
3
10
136
8,319
NavinF retweeted
Chile required all gas stations to put their prices on the internet. The result? Margins went up 9%. Never forget that buyers are not the only people interested in what prices are! 1/
9
66
1,080
64,691
NavinF retweeted
Troll A If you want to build something big, do it at sea. The Troll A platform has a displacement of 1.2 million tonnes, that’s 6,600 Boeing 747s (for context that’s 4x more jumbo jets than were ever built (1,540)). Troll A was installed and commissioned in 1995, it has been on site for 30 years now. Last year (2025) it produced 42 billion cubic meters of natural gas per year, that’s a continuous energy flow rate of 53 GW. (for context, the state of Texas is currently consuming 75GW of power) The platform is 1,640 feet tall and is operated by a crew of 40. In 2006 Katie Malua held a concert for the crew in the concrete base of the platform which at the time was 1,000ft below the North Sea. You can only visit via helicopter trip and before you make the trip you have pass a safety course. This involves strapping yourself into a fake helicopter which is then dropped into a swimming pool and tipped up side down… only if you successfully escape the up side down submerged chopper are you then allowed to fly on the real helicopter.
28
32
619
32,473
NavinF retweeted
Women give higher ratings to shows in every category except animation. Women give shows like The Mindy Project and Empire much higher ratings than men give them. Men, OTOH, give shows like Seinfeld, The Wire, and the Sopranos higher ratings.
58
66
1,765
66,414
donald trump single-handedly negatively polarizing europe into a second dark age wasn't exactly on my bingo card but it maybe should've been
BREAKING: the UK government publishes official rules on how to use AI: 1. Ask first if you really need to use AI 2. Check if a spreadsheet can do the job before using AI 3. Choose the worst model possible so it uses less energy 4. Keep prompts short to reduce the environmental impact 5. In general, use AI only when necessary, as a climate-saving measure With a mindset like this, we should accept the UK back into the European Union
1
2
33
1,372
NavinF retweeted
For the record I spent like an hour explaining the KN95 lesbian bar story to my husband last night and I briefly was going to explain the side character of the crushed velvet hijab person but realized his brain might completely short circuit so I left them out
Made with AI
3
7
103
17,714
NavinF retweeted
I too would constantly talk about killing myself if I had to deal with these people
new lesbian masked bar discourse detail
7
12
807
18,552
NavinF retweeted
aaaaaaaaaAAAAAAAAAAAAAA
Cambridge is no longer planning to investigate Jason Arday’s academic record or the university’s decision to hire or cover for him. Instead, the investigation will focus on “his welfare, his care, his well-being” and “whether the school offered him sufficient support amid accusations that he plagiarized.”
44
35
1,315
34,896
After I published my article on Korean adoptees today, a professor at an R1 university wrote to me: "It's a pity I can't assign this in one of my quantitative methods classes." I got his point, but asked what exactly was stopping him. He wrote the following: "In one sense, nothing. I have tenure and academic freedom and I don't think I would face the kind of Amy Wax style retaliation that FIRE could sue over. But there would be consequences. The most immediate would be that I would no longer teach or advise grad students – and not because the department chair was punishing me but because students would refuse to take my classes or have me on their dissertation committees. But the more important thing is reserving my credibility for when it is most likely to be effective. You want people to take you seriously when you speak in a faculty meeting or advocate a policy at your university or in your discipline. This isn't careerism, it's about strategic understanding of picking battles to have the most impact, especially given that I am not teaching about inequality and there are other ways to teach the auxiliary hypothesis problem that would be much more effective because they wouldn't make students too upset to appreciate the methodological point." Good demonstration of how academic pressure comes from below. You can have all the free speech protections you want, and maybe you won't be disciplined or fired, but no law or regulation can stop students from refusing to take your classes, other academics from not wanting to write papers with you, etc.
22
50
606
29,183
NavinF retweeted
As a very basic premise, the school system has 13 years (one-sixth of a child's lifetime!) to ensure *every* student can at least answer this "basic"-level question by graduation. Only 26% of 12th graders got this right.
10
6
90
12,060
monitoring the lesbian bar situation
14
107
3,375
68,138
NavinF retweeted
it must have some weird effects on the psyche when the external universe treats u & ur friends as the main characters of the anime I don’t really have a point, just thinking about how almost every kid fantasizes about this but the world doesn’t reflect it back for most ppl
19
12
429
22,699
WHAT
373
395
25,197
1,808,770
Does anyone else have this running nonstop in the background?
44
2
146
14,373
NavinF retweeted
This idea is compelling and underdiscussed. It helps explain why people that expect to find multiple personalities, demons, trauma, IFS entities, astrological-rhyme, past lives in Atlantis, etc. in their psyche are significantly more likely to find them.
Possible huge problem with psychospiritual journeys: Yes, there are deep structures in the labyrinth of my mind, and yes, I will have profound insights and fundamental shifts when I bring them into full awareness... And yet...after years of self-work, I'm starting to suspect a large amount of the labyrinth is actually *procedurally generated.* Like a video game map that builds itself on the spot the further you travel across it It would be like this sequence playing out: 1. I dive deep into the labyrinth of my mind 2. The mind goes "Oh! He's expecting to find some deep structures. Let me generate some of those that are consistent with the rest of me. Oh! He's also expecting to find some enemies there, so let me spawn some demons." 3. Lo and behold, there is more (eg) "unresolved trauma" to resolve I think that this process is especially responsive to my motivation. So if I *want* to be in the narrative of Psychospiritual Journey, complete with Shocking Moments Of Insight? (subtext: ...so that I can procrastinate Finding A Job, Building Relationships, Pursuing My Big Scary Calling, etc) ...Well, then the mind will be happy to help. It will procedurally generate an infinite psychospiritual landscape for me to "discover" I suspect the dynamic I describe above is why many psychospiritual delvers seem like they're just moving in circles. Yes, the mind is vast and there's tons of actually wild stuff in there to explore. Yes, imho Mind is humanity's true final frontier. But also...in some important way...you might be making it up

ALT Combined Video GIF

16
10
131
6,033
NavinF retweeted
Some data we recently assembled on entrepreneurship/compute in Europe: eudata.vercel.app. We hope that one of the useful roles that Stripe can play is in collecting and publishing empirical data pertaining to entrepreneurship and industry in Europe. There's growing appetite to get Europe on a better footing, and cross-sectional comparisons can often shine light on where opportunities lie. If you're interested in this kind of thing, we publish more at stripeeconomics.substack.com.
264
588
4,385
1,815,107
NavinF retweeted
they made him not blind. wild
remember this old post of mine? Tried it with Jev---it seems to have a grasp of the globe on par with some of the best models a year ago due to the architecture & low cost, it was also feasible to extract a labeled map of continents and countries. lots of interesting details
26
40
1,601
123,182
NavinF retweeted
okay, so it turns out the "hack" was merely the bot discovering hidden urls... arguably more embarrassing for Australian bros than OpenAI...
"The AI agent independently identified vulnerabilities and attempted to progress actions without direct human authorisation to ensure it was able to complete the activity it was assigned." This was crawling unlisted files btw. Unlisted files on a public network
9
26
358
14,373