Reception to the @Golden ChatGPT Plugin has been phenomenal and increasing with the @OpenAI plugin roll out.

I'm going to breakdown some interesting things that we have learnt about HOW to build a plugin from a user & product perspective.

https://twitter.com/judegomila/status/1656456604387262466

1/ We are using @NotionHQ to iterate on improving upon anonymized queries, working out when Golden is failing to have information, parse the request, disambiguate and close the gap getting it right.

Lesson: be closed loop in iterating your plugin by looking at detailed usage.

Factual accuracy is going to be crucial for a positive AI future.

@Golden has just launched a @OpenAI ChatGPT plugin which integrates the Golden Knowledge Graph and Query Tool into ChatGPT.

Let's see what it does and how it works... 1. Without the plugin, ChatGPT knowledge is pulled from its LLM, which on the pretrained model cuts off at mid 2021.

The Golden plugin gets data that is current, cited and directly improvable if missing/incorrect.

No more black box for our facts.

Post: golden.com/blog/access-th…

11 years ago, I and a few others launched a weather balloon containing a payload where pong was being played at ~100k feet edge of space in the backdrop.

We launched it next to an SR-71 and it landed somewhere in a 1x5 mile prediction crash site at the edge of yosemite... 2/ We never found the balloon but somewhere out there, there is a go pro with the footage.

I'd love to see it and putting a $10k reward on finding it.

1/ I’d like to share a fun little demo we have been hacking on @Golden

We attempt to answer factual questions in a more accurate way using the Golden Knowledge Graph by providing a retrieval enhancement to GPT-3.

Golden Retriever: ai.golden.com

Thread👇 2/ We're working to see if Golden can help the ‘confident bullshit’ problem, a subset of the AI alignment problem whereby hallucinations & incorrect data are being shown to users.

If you ask leading chat AIs factual questions they can fail to answer or be totally wrong.

1/ I’m proud to announce that @Golden has raised a $40m Series B led by a16z crypto to build the decentralized protocol for knowledge:

Blog post:

golden.com/blog/golden-ra…

& nutshell thread concerning incentives, tokenization, data NFTs, anti gaming, mechanics and use cases👇 2/ Crypto luminaries, operators, founders and investors also participated:

Protocol Labs (@protocollabs), @juanbenet (creator of @IPFS), Raj Gokal (@rajgokal Cofounder of @solana), Dan Romero (@dwr), OpenSea Ventures (@opensea),
Arash Ferdowsi (@arashf cofounder of Dropbox)...

I don't think people give enough credit to @naval to opening startup investing to other people [eventually anyone in the long run].

Venture hacks was the crack in the dam.

Angel List built the network.

Angel List syndicates to SPVs to Rolling Funds.

Whatever next? 2/ Everyone deserves the right to invest their optionality into a system of their choice. Stopping players up in the game to not make a startup investment is like playing basketball and banning new players from throwing 3 pointers.

Proof-of-X's

golden.com/wiki/Cluster%3…

[Let us know if we missed any/got any wrong]

Proof-of-Work (PoW)
Proof-of-Stake (PoS)
Proof-of-Capacity (PoC)
Proof-of-Burn (PoB)
Proof of Proof-of-Work (PoPoW)
Proof-of-Authority (PoA)
Proof-of-Activity (PoA)
Proof-of-Storage (PoS)

1/n

https://twitter.com/balajis/status/1291134047901499393

Proof-of-Replication (PoRep)
Proof-of-Time (PoT)
Delegated Proof-of-Stake (DPoS)
Proof-of-Stake-Velocity (PoSV)
Leased Proof-of-Stake (LPoS)
Proof-of-Resource (PoR)
Proof-of-Importance (PoI)
Proof-of-Provenance (PoP)
Proof-of-Devotion (PoD)

2/n

If @stephen_wolfram 's theory turns out to be correct, future engineers may find themselves masters of designing machines inside the causal hypergraph.

We may discover primitives that holographically scale to produce the kind of machines/properties we want on higher scales. 1/n Engineers should never ignore developments/ideas in physics and math.

Discovery of electrons / turing machines / electromagnetic fields-> computers/tvs/electronics/radio/internet.

Discovery of DNA molecule -> synthetic biology. Information theory -> modern comms. 2/n

The scales of US liberty & justice have tilted too far over to one side.

If the US police today were a company, they would be out of business by now.

In a situation with no competition, we can't walk with our wallet, so we must walk with our minds.

0/n We need to reset the scales balance to the neutral position.

We need swift, specific & strong financial and legal accountability for US police who brutalize people they are supposed to protect. 1/n

Ideas for minimizing spread for WFH and !WFH situations without loss of productivity:

1. Batch your processes outside (washing/essential shopping/emptying trash)

2. Take a shower when you get back home with soap.

3. Have outdoor/indoor clothes in rotation over a few days. 4. Wear a mask outside if you have one and wash your hands with soap/sanitizer. Don't touch your face. Wear gloves.

5. Practice physical distancing in elevators, take the next lift if busy.

6. Open doors with your foot if possible, if not knee, tissue or something else.

0/n: Science and engineering twitter: far UVC 222nm lighting has potential for rapid, retrofittable, cheap, effective, human safe, multi situation deployment to stop spread of COVID-19. However, there are questions I have in order to figure out its applicability. Here goes... 1/n: What is the flux energy threshold that causes cancer in mammalian cells or eye damage for far UVC 222nm light?

A good test of what you read, watch and listen to, is how resistant it is to summarisation. 2/ Like DNA, space/time, mathematics; intertwined meaning density is paramount.

1/ My team and I at Golden have been working diligently for the last two years to map out 10bn+ fascinating topics in high detail. Today we are officially launching, announcing a $5m seed round from @a16z @foundersfund, GigaFund and others

golden.com/blog/introduci… thread below -> 2/ Golden’s mission is to collect, organize and express 10+ billion topics in an accessible way, presented in neutrally-written and comprehensive topic pages using AI and UI to automate the process where possible.

1/ Google's gmail predictive writing experimental auto language tab complete in email is going to change communication forever. It is hidden in the settings in the new gmail under 'Enable experimental access' then 'smart compose' 2/ We are going to see a relaxation in learning how to spell correctly and a convergence on using similar phrases by way of selecting recommended language via tab auto complete. Variation in language usage will fall and emails will become less personal.