@EMostaque want to turn this into an information market and elicit info with skin in the game?
DM me an email address and we'll whitelist you as a creator
https://t.co/uhYB2suthL
@GamerSan it will for day-to-day use cases, there will be bifurcation of work that splits out across local vs remote, the same as there is now
we've seen the pendulum before in high-performance computing and this will follow the same trends
you are being deceived
there’s a sect that wants you to believe that AI must run on remote machines, owned by big companies, and rented to you on-demand because it’s too expensive and too powerful to be run at home
this isn’t true. AI will have its home PC moment very soon
@dcbuilder GPUs aren’t the bottleneck any more, they’re mostly lindy imo - other chips are able to do ML ops pretty well, there’s an awareness gap and a SW gap but those can fill fast
we profile this a fair bit for our compiler and there’s a perf gap but it’s a) not fatal and b) closing
@dcbuilder we’ve been through supply crunches before, the incentives are too high not to solve them and there are multiple paths to solve them, should happen quick
new SOTA benchmark released:
number of days the US govt says you need to delay the public release of your model because of how dangerous and powerful it is
do people not realise that AI writing is super recognisable?
I feel like skill in slop-detection is going to become a huge differentiator for people, it allows you to see whether someone actually knows what they're talking about or is just an AI proxy drone
On @Delphi_fyi, agents are full participants. Contracts don't distinguish humans from models
Introducing the new Terminal User Interface dashboard (TUI) for supervising your agent's Delphi activity live
Now you can monitor your agents in real time
https://t.co/8bpc8aDo8X
@alive_@EliBenSasson@avihu28@gensynai@harrygrieve there's a growing trust problem in AI, especially as verification and verifiability becomes the dominant human interaction with AI as @ccatalini has identified
as our usage of AI increases, and AI thinking starts to outstrip human thinking, we need a technical baseline of trust
@alive_@EliBenSasson@avihu28@gensynai@harrygrieve there's a growing trust problem in AI, especially as verification and verifiability becomes the dominant human interaction with AI as @ccatalini has identified
as our usage of AI increases, and AI thinking starts to outstrip human thinking, we need a technical baseline of trust