https://github.com/disarticulate/y-webrtc/blob/master/src/y-... has a validMessage function passed into the room. This allows you to validate any update and reject them. It might be "costly", but it lets you inspect the next object. Since Yjs doesn't care about order or operations, it doesn't really matter how long validation takes.
Not sure what the error conditions loop like, but you could probably bootstrap message hashes in a metadata array in the object, along with encryption signatures to prevent unwanted updates to objects.
Is this an ASIC? Or FPGA? Or something even more exotic?
I’m guessing it’s some form of ASIC because I can’t imagine crafting the logic of Llama on silicon is a very quick or easy job. Not that doing it on an ASIC is a piece of cake either.
"Taalas is borrowing some ideas from the structured ASICs of the early 2000s to make its hardwired model-specific chips. Structured ASICs used gate arrays and hardened IP blocks, changing only the interconnect layers to adapt the chip to a specific workload. At the time, this was seen as a more cost-effective alternative to a full-custom ASIC that was more performant than an FPGA."
"Taalas changes only two masks to customize a chip for a specific model, but the two masks can change both model weights and dataflow through the chip. On the HC1, the model and its weights are stored on the chip using a mask-ROM-based recall fabric paired with a (programmable) SRAM, which can be used to hold fine-tuned weights and/or the KV cache. Future generations of chips may split the SRAM onto a separate chip, meaning they could be denser than the HC1."
One would think that LoRAs being so successful in StableDiffusion, that more people would be focused on constructing framework based LoRas; but the economics of all this probably preclude trying to go niche in any direction and just keep building the do-all models.
The SD ecosystem in large part was grassroots and focused on nsfw. I think current LLM companies would have a hard time getting that to happen due to their safety stuff.
Fine-tuning does exist on the major model providers, and presumably already uses LoRA. (Not sure though.)
We saw last year that it's remarkably easy to bypass safety filters by fine-tuning GPT, even when the fine-tuning seems innocuous. e.g. the paper about security research finetuning (getting the model to add vulnerabilities) producing misaligned outputs in other areas. It seems like it flipped some kind of global evil neuron. (Maybe they can freeze that one during finetuning? haha)
I'm surprised there's no more attempts to stablize around a base model, like in stable diffusion, then augment those models with LoRas for various frameworks, and other routine patterns. There's so much going into trying to build these omnimodels, when the technology is there to mold the models into more useful paradigms around frameworks and coding patterns.
Especially, now that we do have models that can search through code bases.
yeah, you'd think these commercial organizations would sit down with like, one marketer, and just put a non-trivial app together in real time and screen cap it all...
like, we've had this technology for several decades now, and none of these AI tools are like: "This is so great, let me show everyone how to write a CRUD database with a notepad and calendar app" or whatever.
Several decades ago, we barely had the internet, rockets were single use only, and smart phones were coming any day now. CRISPR was yet to be named, social media meant movies from Blockbusters or HBO that you watched with friends. GLP-1 was a meh option for diabetics.
I agree with your overall point but...your time frame is way off.
It's probably more to do with the intelligence required to know when a specific type of code will yield poor future coding integrations and large scale implementation.
It's pretty clear that people think greenfield projects can constantly be slopified and that AI will always be able to dig them another logical connection, so it doesn't matter which abstraction the AI chose this time; it can always be better.
This is akin to people who think we can just keep using oil to fuel technological growth because it'll some how improve the ability of technology to solve climate problems.
It's akin to the techno capitalist cult of "effective altruism" that assumes there's no way you could f'up the world that you can't fix with "good deeds"
There's a lot of hidden context in evaluating the output of LLMs, and if you're just looking at todays success, you'll come away with a much different view that if you're looking at next year's.
Optimism is only then, in this case, that you believe the AI will keep getting more powerful that it'll always clean up todays mess.
I call this techno magic, indistinguishable from religious 'optimism'
Not sure what the error conditions loop like, but you could probably bootstrap message hashes in a metadata array in the object, along with encryption signatures to prevent unwanted updates to objects.
reply