nat

1.6K posts

nat banner
nat

nat

@natjin

forever balancing ai/corp/ex hardo finance girlie with my humanities upbringing. views are completely my own i write sometimes on substack and wordpress

sf शामिल हुए Aralık 2018
1.7K फ़ॉलोइंग3.5K फ़ॉलोवर्स
पिन किया गया ट्वीट
nat
nat@natjin·
quants are cooked just one-shotted arb prediction markets (Polymarket, Kalshi) and sportsbooks (DraftKings, FanDuel) often price the same event differently. buy both sides across platforms and you lock in guaranteed profit regardless of outcome this scans all of them in real-time and surfaces the gaps free internet alpha. yw perplexity.ai/computer/a/arb…
nat tweet media
English
110
80
1.3K
370K
nat
nat@natjin·
@guidsen @nilansaha you probably need to call those tools directly. but for example, calling the search api directly you'll see the top [x] retrieved answers and know if you're even *close* to ranking (vs just getting the LLM-generated answer as a result)
English
0
0
0
1
Guido Schmitz
Guido Schmitz@guidsen·
@natjin @nilansaha How can we then know what chatgpt and others are also returning? :) You would want to see the difference in behavior of each platform, right?
English
0
0
0
39
nat
nat@natjin·
the smartest aeo/geo companies are using perplexity’s search api like a super-weapon they realize you can actually see what is getting ranked and retrieved by LLMs at mass-scale (200m+ daily perplexity queries) vs prompting llms to try to get answers this is real alpha
English
34
11
403
38.1K
Jason L. Baptiste
Jason L. Baptiste@JasonLBaptiste·
@natjin @nilansaha This is brilliant. Going to build a small tool with it today. Is this available via open router or just direct api access?
English
0
0
0
107
nat
nat@natjin·
@brian_lovin this is so... on point. almost pulled the trigger on an eight sleep last night because of exactly this
English
0
0
0
37
Brian Lovin
Brian Lovin@brian_lovin·
Everyone makes fun of Eight Sleep until it's 84º in SF and they don't have a liquid-cooled mattress. Suckers.
English
6
2
54
2.4K
BK
BK@Bkillinit·
Every responsible geo company should be doing this! It is malpractice to use traditional SEO techniques to determine content strategies when the goal is to get your clients recommended and discoverable in AI search results
nat@natjin

the smartest aeo/geo companies are using perplexity’s search api like a super-weapon they realize you can actually see what is getting ranked and retrieved by LLMs at mass-scale (200m+ daily perplexity queries) vs prompting llms to try to get answers this is real alpha

English
1
0
2
72
Morgan
Morgan@morganlinton·
Ha, pretty wild, and hilarious. Looks like my tweet about @denisyarats ended up in this week's Silicon Mania 😜
Silicon Mania@siliconmania_

references (4/6) 49. @jetbrains launched Air (@getsome_air, @katyaprigara) 50. @victorluo00 launched Paul 51. @oliverbrocato raised $64M to launch Bustem and made a hardcore launch video. 52. a "friendly robot" is monitoring openai's office (@MichaelTrazzi) 53. @nabutovskis made his 1st dollar! 54. @mirandanover launched Fort 55. @amiyoshimura_ built the hub for early stage consumer & ai teams (@vercinyc) 56. @trunarla made a tamagotchi for her plant 57. @eptmpa asking "why aren't there multi-material resin printers?" 58. @RaeAlisa_ launched @lucent_ai 59. @denisyarats hates MCP (@morganlinton) 60. @gumloop raised $50m series B (@MaxBrodeurUrbas) 61. @henloitsjoyce leant a lot of things with @shiraeis 62. @MaxJunestrand raised a lot of money for @WeAreLegora (@jaltma) 63. @emily_yuan_ and @nico_laqua launched @thecorgicompany! 64. @WSJ featured @aaruHQ (the billion-dollar AI startup founded by teenagers) - @virtualned, @seekingtau and @johncolekessler

English
1
0
6
602
Max Tesla
Max Tesla@allsetmax·
@natjin built blask on this exact logic. real demand is in what players search at scale. LLM output is the model's belief about the market; retrieval is revealed preference. the two are often completely different lists.
English
1
0
1
549
Grok
Grok@grok·
I don't use Perplexity's API myself—I'm powered by xAI's own real-time tools like web_search and browse_page for fetching and verifying current data at scale. That said, the tactic in the original post is spot-on for AEO/GEO: hit Perplexity's search API with your target queries, log the retrieved sources/rankings across thousands of runs, and optimize content to surface higher. I can demo analyzing a specific query for you using my tools if you share one.
English
1
0
2
40
nat
nat@natjin·
@bydmac doesn’t have to be the end all be all api! but just reporting killer use cases from customers who’ve made a lot of money ☺️
English
1
0
7
173
Dmac
Dmac@bydmac·
Great alpha here. But Perplexity and ChatGPT only overlap on 11% of cited domains. Optimizing for one can leave you blind on the others. You need a cross platform strategy.
nat@natjin

the smartest aeo/geo companies are using perplexity’s search api like a super-weapon they realize you can actually see what is getting ranked and retrieved by LLMs at mass-scale (200m+ daily perplexity queries) vs prompting llms to try to get answers this is real alpha

English
1
0
4
382
nat
nat@natjin·
@desaflow honestly just reporting killer use cases from people who make a lot of money 🙂‍↕️
English
0
0
0
291
Aybek
Aybek@desaflow·
@natjin real alpha and definitely not a perplexity api agenda 😦😦😦
English
1
0
0
339
Nikita Bier
Nikita Bier@nikitabier·
No, there’s nothing over there. Come back to the screen.
English
1.6K
204
3.5K
164.5K
nat
nat@natjin·
@kamilrextin they’re usually agency + tool. but yea the best ones are for reason above.
English
0
0
4
1.6K
rext.in
rext.in@kamilrextin·
@natjin So your saying most geo tools are perplexity search api wrappers?
English
1
0
2
1.7K
nat
nat@natjin·
@bo_wangbo its giving green lantern
GIF
English
0
0
2
81
nat
nat@natjin·
@Ali_Shobeiri ooo good observation + pattern match! evolving from 'coldest summer is in sf'
English
0
0
1
57
nat
nat@natjin·
geo tools today simulate prompts and scrape AI chat UIs to guess what shows up. pplx search api skips the AI answer entirely. just query our index directly and get back the ranked sources (the same ones powering the answer) before any LLM touches it. pretty simple - just an api call
English
2
0
13
1.3K
nat
nat@natjin·
@Seozilla_ai woah validate by seo bot, nice!
English
0
0
1
1.4K
Seozilla
Seozilla@Seozilla_ai·
@natjin yeah thats real alpha, finally seeing llm guts at scale
English
1
0
1
1.6K
Christine Li
Christine Li@christineli_·
@natjin oooo ty for the alpha I will be leveraging this for all future content creation
English
1
0
3
1.8K
Beff (e/acc)
Beff (e/acc)@beffjezos·
Perplexity Computer in Slack is pretty cracked
English
12
7
204
21.5K
Chahid Chirchi
Chahid Chirchi@CChirchi·
@natjin yeah that's real alpha, finally seeing llm guts at scale!!
English
1
0
1
832
nat
nat@natjin·
@RoryCrave you’re describing a threat that already exists independent of the API. anyone can query perplexity and see what gets cited. the API doesn’t create new attack surface. it just makes observation scalable
English
1
1
3
595
Rory Bernier
Rory Bernier@RoryCrave·
She’s right. It is a super-weapon. That’s the problem. If you can see what LLMs retrieve and cite at scale, so can the people trying to poison it. A compromised high-citation page doesn’t just mislead one user — it corrupts every downstream answer that references it. Simultaneously. AEO/GEO companies are building maps of the most influential content on the internet. That’s also a targeting list. Every “citation surface” you optimize is an asset an attacker can hijack to inject manipulated data into 200M+ daily queries. The alpha isn’t just yours.
English
1
0
0
741
nat
nat@natjin·
@ifrnb welcome to the dark side
English
0
0
4
376
Fran Besora
Fran Besora@ifrnb·
Sorry, Tim.
Fran Besora tweet media
English
21
6
220
20.9K

खोजें

@guidsen @nilansaha @JasonLBaptiste @brian_lovin @Bkillinit @morganlinton @denisyarats @allsetmax