Biggest AI surprise will be if it ends up being AI boyfriends, not AI girlfriends, that really takes off
Guys usually more physical, women more non-physical. Would make sense. Hello Great Filter.
Guys usually more physical, women more non-physical. Would make sense. Hello Great Filter.
👀8👍5🤯2❤1😱1
GPT-4 for content moderation: ultimate censorship tool
FWIW: OpenAI suggests, for speed and cost reasons, the architechture of:
(1) Collect a balanced set of example content to moderate + moderation instructions, and then
(2) Get GPT4 to auto-label that set as good or bad, and then
(3) Train a fast small classifier on this dataset
I.e. just use GPT4 for generation of training labels, not for live moderation.
Reminds me a bit of Snorkel: Rapid Training Data Creation with Weak Supervision
Next gen censorship tools gonna be lit.
OpenAI: Using GPT-4 for content moderation
FWIW: OpenAI suggests, for speed and cost reasons, the architechture of:
(1) Collect a balanced set of example content to moderate + moderation instructions, and then
(2) Get GPT4 to auto-label that set as good or bad, and then
(3) Train a fast small classifier on this dataset
I.e. just use GPT4 for generation of training labels, not for live moderation.
Reminds me a bit of Snorkel: Rapid Training Data Creation with Weak Supervision
Next gen censorship tools gonna be lit.
OpenAI: Using GPT-4 for content moderation
🤬6❤3🔥2⚡1😁1😱1👀1
Open challenges in LLM research
“1. Reduce and measure hallucinations
2. Optimize context length and context construction
3. Incorporate other data modalities
4. Make LLMs faster and cheaper
5. Design a new model architecture
6. Develop GPU alternatives
7. Make agents usable
8. Improve learning from human preference
9. Improve the efficiency of the chat interface
10. Build LLMs for non-English languages”
He’s correct that #1 and #2 are among the top priorities, both far more non-trivial than they look, and for both, the existing naive proposed solutions are obviously dead wrong and irreparable.
Article
“1. Reduce and measure hallucinations
2. Optimize context length and context construction
3. Incorporate other data modalities
4. Make LLMs faster and cheaper
5. Design a new model architecture
6. Develop GPU alternatives
7. Make agents usable
8. Improve learning from human preference
9. Improve the efficiency of the chat interface
10. Build LLMs for non-English languages”
He’s correct that #1 and #2 are among the top priorities, both far more non-trivial than they look, and for both, the existing naive proposed solutions are obviously dead wrong and irreparable.
Article
👍7❤1
Gartner over here trying to convince us with their kindergartner scribbles
Looks like it’s Gartner, whose BS is past its peak
(Like stock-to-flow, whose only legitimate idea is that BTC price roughly continues on an exponential attenuating curve, and the rest of the details all BS — Only legit part of Gartner’s theory is that sometimes legit things do go through booms and crashes, but every other detail beyond than that, eg the shape of the curve, number of crashes, idea that any big hype always results in a big crash — all BS, not matching reality.)
Looks like it’s Gartner, whose BS is past its peak
(Like stock-to-flow, whose only legitimate idea is that BTC price roughly continues on an exponential attenuating curve, and the rest of the details all BS — Only legit part of Gartner’s theory is that sometimes legit things do go through booms and crashes, but every other detail beyond than that, eg the shape of the curve, number of crashes, idea that any big hype always results in a big crash — all BS, not matching reality.)
👍4❤1
Gartner Hype Cycle -vs- Reality
Legit productivity tech that’s locked into the legit path and creating real value essentially never experiences an severe Gartner-shaped adoption or financing crash.
What does?
Legit fields that get subverted and redirected onto a different scam, non-value-creating path.
E.g. Earlier AI waves, that got redirected from the legit Connectionist path, onto the scammy pure-Symbolic path — something that never got to the point of creating real AI value for companies.
E.g. Current crypto, that got redirected from the path of 100% identity-irrelevant trustworthiness with 0% reputation-based trust, to 100% identity and reputation based Worldcoin scams. Some portion creating legit value for criminal organizations, but not big enough of a portion.
Gartner-like cycles happen when legit fields get redirected into pure scam territory.
AI? It’s finally gotten on the legit, actual value-creating path. Point of no return.
Welcome to the last AI summer, the summer that never ends, like electricity, phone, email.
Legit productivity tech that’s locked into the legit path and creating real value essentially never experiences an severe Gartner-shaped adoption or financing crash.
What does?
Legit fields that get subverted and redirected onto a different scam, non-value-creating path.
E.g. Earlier AI waves, that got redirected from the legit Connectionist path, onto the scammy pure-Symbolic path — something that never got to the point of creating real AI value for companies.
E.g. Current crypto, that got redirected from the path of 100% identity-irrelevant trustworthiness with 0% reputation-based trust, to 100% identity and reputation based Worldcoin scams. Some portion creating legit value for criminal organizations, but not big enough of a portion.
Gartner-like cycles happen when legit fields get redirected into pure scam territory.
AI? It’s finally gotten on the legit, actual value-creating path. Point of no return.
Welcome to the last AI summer, the summer that never ends, like electricity, phone, email.
👍3❤2👏1