the masculine urge to buy an old apple ii and program little games and animations on it in your spare time
@jxmnop
-
Career reputation risks in technology advocacy and idea promotion
By
–
(i've also heard that you can get a reputation for doing this sort of thing and it doesn't pay off in the long run of a career promoting your ideas)
-
Truth Over Self in Research: Why Academics Avoid Self-Promotion
By
–
but I think the reason why people don't do this more is that good researchers are concerned with presenting the *truth*, putting the importance of the ideas above the importance of the self
-
Scientific Publishing Integrity and Deceptive Result Presentation Risks
By
–
a weird quirk about publishing in science is that it's really easy to be deceptive in your presentation of results. can be trivial to find loopholes in a complex system that only you fully understand
-
Superhuman AI Capabilities Discussion
By
–
don't think so, at least the reason why I chose this example is that it feels superhuman, or at least only within the capabilities of a very small group of humans
-
Will larger language models eventually enable critical infrastructure attacks?
By
–
Imagine this prompt: "write me Python code to disable the NYC subway system" obviously gpt-4 can't do this now. It'll refuse, but even if we jailbreak it, it'll answer incorrectly but if we keep training bigger & better language models, won't one eventually be able to do this?
-
Simultaneous Discovery of Task and Function Vectors in AI
By
–
simultaneous discovery of “task vectors” and “function vectors” is another interesting case in a long history of different people in different places having the same idea at the same time just like calculus, the telephone, prefix tuning…
-
Data Value and ChatGPT Pricing Models
By
–
guess it really depends on (i) the value of this type of data and (ii) chatGPT pricing
-
Scientists Collected 570K ChatGPT Conversations for API Access
By
–
these mad scientists really just allowed strangers to use their chatGPT api key for free in return they collected 570k real conversations with chatGPT I’ve heard this data is super valuable, but was this really worth it?
-
Distributed GPU Processing: File Synchronization and Aggregation
By
–
because in distributed mode, each GPU runs its own process. so this code will run simultaneously multiple times, once per GPU. so in this setting, each GPU writes a file, waits for everyone, then aggregates all the files, waits again, then deletes its own file 🙂