This is true, but consider that we have agentic token injection (what you are referring, prompting), embedding injection (memorilla) and continual learning.
So the first will update _what_ information the model reasons/infers about, the last will update _how_ the model reasons/ infers about.
These are very different.
As far as I understand memorilla sits in the middle. And it embeds not tokens but embedding vectors with all their complexity. The compression (semantic) is significant and in my mind is super interesting.
What do you think about the efforts of Hiley and Denis [1], which may suggest that the pilot wave may be a manifestation, a mathematical projection, of a deeper physical process which hierarchically gives way to the pilotwave, particles, like a coordinati9n space.
They were talking about modeling orbits with Keplerian gravity and I mentioned that GrC(2,4) does seem to fit QED scattering gravity also with zero error.
This is the extent of the peer review this far: https://news.ycombinator.com/item?id=50289590#50295549
Skimmed the paper briefly. Looks like reformulation of single-particle QM with no new predictions. Most importantly (and as I mentioned above) doesn’t address particle creation and annihilation, which is where Bohmian-style approaches usually struggle.
Authors seem to be aware of that, but make a weird tap-dance around the problem throughout the paper, seemingly scared to even "Voldmort" it. In that context, citing 1930 Dirac paper is laughably ironic.
One can argue it gives better ontology of what is going on, but shifts the burden further into what the pilot wave is. And this actually is far more complicated maybe than QM itself.
Transactional interpretation of quantum mechanics is compatible with relativity and it is it that deserves more attention than Bohm interpretation. Unfortunately it uses a terminology like retrocasuation which sounds even stranger than how Copenhagen interpretation explains QM. So as long as experimental predictions stays the same I see little chances that transactional or similar interpretations will offset Copenhagen mainstream.
Why not work as before (on 3D XPoint) with Micron and try to pursue PIM (processing in memory) on DRAM? It doesn't go around the DRAM manufacturing constraint, but it helps them sell new PIM-DRAM coupled to Xeons. Right now, they have "pie in the sky" GPUs and nVidia has their boot on Intel's throat, while Xeons and let alone normal cpus are going down the drain.
Intel needs to innovate under constraints and, if history is to be reinterpreted, this won't end well for them. They need less management and to double-down on engineering and their lithography pipeline.
Bold move. Although, I don't think "bold" is what they need where they are going. What's up with these guys? You have something and you wanna turn it into nothing. Just take the site down it will be far less painful.
Aren't shareholders interested in how RDDT is doing?
I was wondering this, too. The desktop appears to be based on herdr-core, and naturally I'd have imagined Herdr Server (terminal) to migrate to herdr-core as well. In fact I'm not even sure why they built herdr-core when they had Herdr Server ..
I was initially excited because I thought I could just try out a new frontend while keeping my 10 existing spaces each with 10 worktrees each with 2 or more tabs untouched.
I fry my brain everyday trying to keep pace, but honestly, stop it guys. Give the guy a break.
This is the most magical time in my adult life. The 70s, 80s, and 90s, look like a painting compared to the current pace of things. I am literally tired at the end of each day. Dog tired.
Your point is valid for the agentic tool search, i am just saying there are two things lurking.
reply