Profile picture
, 3 tweets, 1 min read
My Authors
Read all threads
GPT Trainer

If you can’t afford a personal coach, an AI one will be provided to you free of charge. Select from motivational, empathetic, or provocative modes.

Private personalized feedback from the wisdom of the web.
People confide all kinds of things into Google. But this might be next level. It would almost be like praying to God for advice.

Are you there GPT-3? It’s me, Margaret. goodreads.com/book/show/3773…
There of course have been many different visions of gods over the years. Thor and Jesus would give very different feedback.

So, perhaps it’s user-configurable. A personalized relationship with GPT-3. 😜
Missing some Tweet in this thread? You can try to force a refresh.

Keep Current with balajis.com

Profile picture

Stay in touch and get notified when new unrolls are available from this author!

Read all threads

This Thread may be Removed Anytime!

Twitter may remove this content at anytime, convert it as a PDF, save and print for later use!

Try unrolling a thread yourself!

how to unroll video

1) Follow Thread Reader App on Twitter so you can easily mention us!

2) Go to a Twitter thread (series of Tweets by the same owner) and mention us with a keyword "unroll" @threadreaderapp unroll

You can practice here first or read more on our help page!

More from @balajis see all

Profile picture
balajis.com
@balajis
GenePT-3

A new spin on biomedical literature mining. Take the entire corpus of peer reviewed literature and train a GPT-3-like model on it. Now give in a gene symbol and see what it spits out.

Twist: can we include backlinks so you can see literature used to generate the recs?
You might even type in a biomedical thesis and see what comes up. “We hypothesize that protein X binds to protein Y in organism Z.”

In this context it’s like a new kind of search engine, for generative search.
This is great!

Biomedicine is well suited to this because many relationships between proteins, transcription factor binding sites, small molecules, etc are encoded in natural language across millions of papers.
Read 3 tweets
Profile picture
balajis.com
@balajis
It's time to gradually exit Twitter.

While Substack is great, it's better to set up a paid newsletter at your own domain. You can do this in an hour and maintain it for less than $30 per month.

I wrote up a tutorial here to walk you through it:
balajis.com/set-up-a-paid-…
Why gradually exit Twitter? It's not just the hack. There are many overlapping issues around security, authentication, pseudonymity, distribution, incentives, and culture.

I greatly respect Jack and everything he's built. But we need a change. balajis.com/how-to-gradual…
Feel free to stop after you set up your paid newsletter. It's a useful thing in its own right.

But in terms of what's next, we may be able to knit these newsletters together into a decentralized media community, with individual sites as the new profiles.
balajis.com/how-to-gradual…
Read 9 tweets
Profile picture
balajis.com
@balajis
Meanwhile, in China…

“The local government of Beijing released on Thursday a [145 page] blueprint of its plan to implement a blockchain-based programmable government.”

“Already 140 government services use blockchain, the government said in a statement.”
technode.com/2020/07/16/bei…
This looks real, unlike a lot of vaporware in the space. This is an official government document in Chinese that appears to lay out system diagrams, specific data structures, and API calls.

A translation would be valuable.
beijing.gov.cn/fuwu/lqfw/ztzl…
A programmable government is a powerful thing. A standardization of various APIs, a government as a technology platform. The network state.

I’d need to read the spec, but it may be a hierarchical permissioned ledger where the state has root access but others have read access.
Read 4 tweets

Embed code for your website

×
Follow Us on Twitter!

Did Thread Reader help you today?

Support us! We are indie developers!

This site is made by just two indie developers on a laptop doing marketing, support and development! Read more about the story.

Become a Premium Member ($3.00/month or $30.00/year) and get exclusive features!

Become Premium

Too expensive? Make a small donation by buying us coffee ($5) or help with server cost ($10)

Donate via Paypal Become our Patreon

Thank you for your support!