It’s an interesting question isn’t it? There are obvious qualities about being able to find information quickly and precisely. However, the search becomes much narrower, and what must inevitably result is a homogeneity of outcomes.
Eventually we will have to somehow convince AI of new and better ways of doing things. It’ll be propaganda campaigns waged by humans to convince God to deploy new instructions to her children.
And this outcome will be obvious very quickly for most observers won't it? So, the magic will occur by pushing AI beyond another limit or just have people go back to specialize on what eventually will becoming boring and procedural until AI catches up
I’m finishing up a language identification model that runs on cpu, 70k texts/s single thread, 13mb model artifact and 148 supported languages (though only ~100 have good accuracy).
This is a model trained as static embeddings from the gemma 3 token embeddings.
> Berulis said he and his colleagues grew even more alarmed when they noticed nearly two dozen login attempts from a Russian Internet address (83.149.30,186) that presented valid login credentials for a DOGE employee account
> “Whoever was attempting to log in was using one of the newly created accounts that were used in the other DOGE related activities and it appeared they had the correct username and password due to the authentication flow only stopping them due to our no-out-of-country logins policy activating,” Berulis wrote. “There were more than 20 such attempts, and what is particularly concerning is that many of these login attempts occurred within 15 minutes of the accounts being created by DOGE engineers.”
Every time I see post-DOGE kvetching about foreign governments' hacking attempts, I'm quite bewildered. Guys, it's done, we're fully and thoroughly hacked already. Obviously I don't know if Elon or Big Balls have already given Putin data on all American military personnel, but I do know, that we're always one ketamine trip gone wrong away from such event.
The absolute craziest heist just went in front of our eyes, and everyone collectively shrugged off and moved on, presumably to enjoy spy novels, where the most hidden subversion attempts are getting caught by the cunning agents.
I'm genuinely confused about this story and the affiliated parties. I've actively tried to search for "Daniel Berulis" and couldn't find any results pointing to anything outside the confines of this story. I'm also suspicious of the lack of updates despite the fact that his lawyer, Andrew Bakaj, is a very public figure who just recently commented on a related matter without bringing up Berulis [1].
Meanwhile, the NLRB's acting press secretary denies this ever occurred [2]:
> Tim Bearese, the NLRB's acting press secretary, denied that the agency granted DOGE access to its systems and said DOGE had not requested access to the agency's systems. Bearese said the agency conducted an investigation after Berulis raised his concerns but "determined that no breach of agency systems occurred."
One can make the case that he's lying to protect the NLRB's reputation, but that claim has no more validity than Daniel Berulis himself lying to further his own political interests. Bearese has also been working his position since before the Trump administration started, holding the job since at least 2015. It's very hard for me to treat his account seriously, especially considering the political climate.
Heavens to Betsy,
I don’t know if you can hear me,
But try supporting these things if you actually want them to be successful. About the 3rd day into trying to roll your own LMI container in sagemaker because they haven’t updated the vLLM version in 6 months and you can’t run a regular sagemaker endpoint because of a ridiculous 60s timeout that was determined to be adequate 8 years ago. I can only imagine the hell that awaits the developer that decides to try their custom silicon.
That might not be relevant to OPs use case. A lot of nurses get tied up doing things like reviewing claims denials. There’s good use cases on the administrative side of healthcare that currently require nurse involvement.
I love using encoder models, and they are generally a better technology for this kind of application. But the price of GPU instances is too damn high.
I won’t lie that I’ve been unreasonably annoyed that I have to use a lot more compute than I need, for no other reason than an LLM API exists and it’s good enough in a relatively small throughput application.
One of the issues with using LLMs in content generation is that instruction tuning causes mode collapse. For example, if you ask an LLM to generate a random number between 1 and 10, it might pick something like 7 80% of the time. Base models do not exhibit the same behavior.
“Creative Output” has an entirely different meaning when you start to think about them in the way they actually work.
Creativity is a really ill-defined term, but generally it has a lot more to do with abstract thinking and understanding subtlety and nuance than with mode collapse. Mode collapse affects variation, which is probably a part of of creativity for some definitions of it, but they aren't the same at all.
I wanted a simple retrieval index to use splade sparse vectors. This just encodes and serializes documents into flatbuffers and appends them into shards. Retrieval is just parallel flat scan, optionally with reranking.
The idea is just a simple, portable index for smaller data sizes. I’m targeting high quality hybrid retrieval, for local search, RAG or deep research scenarios.
SPLADE is a really nice “in-between” for semantic and lexical search. There’s bigger and better indexes out there like Faiss or Anserini, but I just kinda wanted something basic.
I was testing it on 120k docs in a simple cli the other day and it’s still as good as any web search experience (in terms of latency) — so I think it’ll be useful.
We’re still trying to clean up the API and do a thorough once over, so I’m not sure I’d recommend trying it yet. Hopefully soon.
I haven’t used RCNN, but trained a custom YOLOv5 model maybe 3-4 years ago and was very happy with the results.
I think people have continued to work on it. There’s no single lab or developer, it mostly appears that the metrics for comparison are usually focused on the speed/MAP plane.
One nice thing is that even with modest hardware, it’s low enough latency to process video in real time.
Eventually we will have to somehow convince AI of new and better ways of doing things. It’ll be propaganda campaigns waged by humans to convince God to deploy new instructions to her children.
reply