@Amirhasan once told me that you cannot have the same feeling twice. And then I read “one cannot pass a river twice for that is not the same river and they are not the same person”
I’m doing a hard work these days, I’m trying to fight systems that are thousands of years in hardening. They have much to say no doubt.
And whenever I—or my team—face hardships, this urge of going back to safety arises. Doing the “right things”. Well that is hard to control, you have to fight.
Dvorak tried to create American music with European philharmonic orchestra, it was stupid of course. It took immense effort and courage for Jazz to appear. And I have to be the one who refuses the safety of the orchestra every single day.
That is one of the hardest things in the world. Even if it does not appear as such.
You see systems and mediums have prejudices. I have made and studied so many that it is my first instinct to smell the prejudices. And so you have to keep systems and routines and procedures out of the room for so long that you can find what is a good system.
Introducing the system too early breaks everything. I have done it so many times. DesignTalk was such a mistake, I studied Maestro Victor for two weeks and started building without a second of thought. Or the Archive . It is hard to live without systems long enough that you can see what a system you actually need.
When you’re building atoms, agile is not your answer. You have to live without systems and without making, and let them in when you’re ready.
Bow I only realize this now. For so long I didn’t know what was happening to me. Why did Pouya Kary who lived all of his life in systems suddenly is resisting them.
And others think I don’t have a system or that I think free. Well that is a mistake. I’m not at all like that. I just don’t know what should be the shape of a LEGO that is my best tradeoff between standards and freedoms.
After a long time I had the chance to come back home and be with Téo . He—after months—came to sleep on my arms, I love my son so very much.
When I’m not at home, I keep thinking about @tea and I hate myself for being an absent father.
Naming is the purest form of graph compression.
And I miss Kamwa .
maintaining a global workspace is likely computationally useful: a common representational format allows intermediate results to be written once and read by many neural processes. A language model that must chain reasoning steps, apply general operations in arbitrary contexts, and answer questions about its own processing also stands to benefit from this organization.
Note that access consciousness is a purely functional notion; the relationship that it has with subjective experience (sometimes called phenomenal consciousness) is widely debated.
Modern large language models (LLMs) are known to perform sophisticated, multi-step internal computations in order to select their actions. As part of their internal processing, might LLMs have developed a global workspace of their own, to serve a functional role analogous to conscious access? It is not obvious that they should; in the brain, the workspace is closely associated with recurrent dynamics and brain region interactions that have no direct analog in the transformer architecture on which LLMs are based. On the other hand, maintaining a global workspace is likely computationally useful: a common representational format allows intermediate results to be written once and read by many neural processes. A language model that must chain reasoning steps, apply general operations in arbitrary contexts, and answer questions about its own processing also stands to benefit from this organization. Even if the implementations differ, it is natural to ask whether the functional properties associated with the global workspace have emerged in LLMs.