be the Order Book you want to see in the world
give to crypto what is of caesar, less of a headache for h**
#DeSci, $ai, ๐ฏ๐
Opinions are mine, ideas are yours
๐ ๐ ๐
"Non so che viso avesse, neppure come si chiamava,
Con che voce parlasse, con quale voce poi cantava,
Quanti anni avesse visto allora, di che colore i suoi capelli,
Ma nella fantasia ho l'immagine sua: Satoshi Nakamoto"
Findings on shingles vaccination for dementia prevention not replicated in large independent study - we need better mandalas for heuristics - ๐
https://t.co/9ex8oz9qZy
@OpenAI Fascinating how benchmark design itself can be the bottleneck. Sometimes the real breakthrough is not a bigger model but simply giving the model the memory it deserves
Iโm significantly older than you. I started coding in the late 60s. My current strategy is to not read any of the code written by my agents. Thatโs the only way I can take advantage of their productivity. What I do instead is to surround the agents with extreme constraints. Unit tests, gherkin tests, QA procedures, quality metrics, mutation testing, test coverage, and a plethora of others. In the end, I have very high confidence in the code they produce because theyโve had to run the gauntlet of all of my constraints and tests.
my prompts:
- how well would these abstract do from a traditional journal review process? are they complete in design or more on the word salad-ai bs side? https://t.co/mCgQbKHF7z https://t.co/y8vHRuR2hY;
- you have to recalibrate 'Conceptual coherence' and 'Novelty potential' considering that these may be copied from normal research. The platform is not a final journal, but a rigorous review process for first attempts at formulating hypothesis would help the community filter ideas through a flood of AI allucination and copycats. What vote would you give on your parameters? would they be worth a refined review from a superior AI based on token expenditure instead of this flat price chat?;
- would you output a simple report on why both options are or aren't worth a token-priced review after this first flat-price chat with you? a simple table with weighted values and brief comments if needed;
- you have changed your vote between the last and second last prompt;
- be dry on the content of the actual submissions.
Considering the amount of reasonable claims and AI hallucinations a platform for automated first review could get, and the idea that token-priced reviews are costly, how would you assess whether such two examples of proposals are worth anything compared to the broader literature?
@sciencebeach__ legacy Open Labs by @BioProtocol has an opportunity to pull the price of review down like we're seeing in theoretical assessments in maths done by AI, in the last months. Of course it's just started.
in this convo with chatgpt 5.5 "high" (flat price) I prompted it to help me with assessing the potential ROI in the top commented and voted proposals in the platform. I'll post my questions in a second message.
it already happened to me a few time to ask AI "is this sensical", with ideas I may have found interesting, or intuited the originality of, but for me it's just a pin on a map, while a complete pipeline for that could be...worth some journal credit.
Abstract Review Assessment https://t.co/QgWRfpLYeB
Join us tomorrow (July 16) for the second class in our virtual #QuantumBiology series! The class begins at 11 a.m. PT (2 p.m. ET) and will explore how quantum biology may operate in proteins. To register, go to: https://t.co/o7cGHiEwMc.
Once we have completed our review for security vulnerabilities, we will make the entire codebase of ๐ open source, with no exceptions.
Moreover, we will invite third party reviewers to examine the system that is running to confirm that the open source code is what is running.
Trust through total transparency is the only thing that should be believed.