HackerTrans
TopNewTrendsCommentsPastAskShowJobs

fertrevino

no profile record

Submissions

From GPT-4 to GPT-5: Measuring progress through MedHELM [pdf]

fertrevino.com
127 points·by fertrevino·11 mesi fa·96 comments

comments

fertrevino
·8 mesi fa·discuss
Trying to get the first commercial users for my digital menu management product: menuop.com

I have been working in ad campaigns and instrumentation to understand interest and reach out to potential users.
fertrevino
·10 mesi fa·discuss
I am rewriting my pet project: menuop.com . I am integrating AI in the whole restaurant menu creation process. Specifically I'm leveraging the Nano Banana to generate suitable images for food items.
fertrevino
·11 mesi fa·discuss
I wonder what exactly would prevent a developer from removing the signature from a generated file. One could remove arbitrary segments that signal that it is AI generated.
fertrevino
·11 mesi fa·discuss
'There seems to be a problem with your account. Please contact Github support' is what I get in my vscode extension.
fertrevino
·11 mesi fa·discuss
loved the cartoon :)
fertrevino
·11 mesi fa·discuss
That would be an interesting extension. MedGemma isn't part of the original benchmark either [1]. Since Gemini 2.0 Flash is on 6th place, expectations are for MedGemma to achieve higher than that :)

[1]https://crfm.stanford.edu/helm/medhelm/latest/#/leaderboard
fertrevino
·11 mesi fa·discuss
Mixed results indeed. While it leads the benchmark in two question types, it falls short in others which results in the overall slight regression.
fertrevino
·11 mesi fa·discuss
Interesting, it seems the anecdotal experience agrees with the benchmark results.
fertrevino
·anno scorso·discuss
I find this topic more relevant as time goes by, I consider compliance and assessment of AI systems is necessary for its own success. I currently conduct a field study on the topic: https://www.ai-compliance.cloud/
fertrevino
·anno scorso·discuss
I am extending my pet project menuop.com, a digital menu maker for restaurants. I plan to integrate AI to enhance visual impact, analytics and recommendations to restaurant owners.
fertrevino
·anno scorso·discuss
I heard russian citizens were having troubles getting any kind of U.S. visa due to the political situation in the world. Does this situation still persists? Asking for a friend who still hasn't heard back for his tourist visa.
fertrevino
·2 anni fa·discuss
I loved this when I started using it. It does work very well to a certain extent. At some point the diagrams became too large and the arrows overlap started to be really inconvenient.

I heard that likec4.dev is a good alternative for this, but I have no hands on experience with it.