motoboi

11 ms·

151.
by motoboi 2y ago
Because hallucinations is something that from a distance looks very unimportant, but when looked closely is a structural problem. Some people here live very close to the LLM field. Structural because while a human being can be the judge of
152.
by motoboi 2y ago
… in a year. Theirs is for 4 years.
153.
by motoboi 2y ago
OpenAI reasoning traces are actually summarized by another model. The reason is that you can (as we are seeing happening now) “distill” the larger model reasoning into smaller models. Had OpenAI shown full traces in o1 answers they would ha
154.
by motoboi 2y ago
I’m my experience and based on writeups like this: Google hates having customers. Someone decided they have to have a public cloud, so they did it, but they want to keep clients away with a 3 meter pole. My AWS account manager is someone I
155.
Repair of tooth enamel by a biomimetic mineralization (2019)(science.org)
4 points by motoboi 2y ago|0 comments
156.
by motoboi 2y ago
JSON rest api are just hypermedia rest api but with the data compressed. The compression format is json and the dictionary are the hypermedia relations previously passed in the client. It’s 2025, the client don’t need to be generic and able
157.
by motoboi 2y ago
Not necessarily. For visual tasks, that is the state of the art, with visual features being "gouped" into more semantically relevant parts ("circles" grouped into "fluffy textures" grouped into "dog ears&q
158.
by motoboi 2y ago
I suppose humans need high level concepts because we can only hold 7[ ] things in working memory. Computers don’t have that limitation. Also, humans cannot iterate over thousands of possibilities in a second, like computers do. And finally,
159.
by motoboi 2y ago
It’s funny and sad to have people talking about types while I have Java and and IDE that basically writes code itself. JavaScript was a bad trip, guys.
160.
by motoboi 2y ago
This is from a model from 2019! Published now on nature. Imagine what they are cooking now.
161.
by motoboi 2y ago
I suppose you didn't get the news, but google developed a LLM that can play chess. And play it at grandmaster level: https://arxiv.org/html/2402.04494v1
162.
by motoboi 2y ago
You are probably familiar with a document called OAuth Threat model. In that document, refresh token rotation is preferred, but it also addresses the obvious difficulty in clustered environments: https://datatracker.ietf.org&#x2F
163.
by motoboi 2y ago
Yeah. Also the young blood always assume the gray haired guys were just some kind of stupid and lame and simple couldn’t see how easy is to setup PHP with Linux on a commodity class hardware running somewhere in the basement.
164.
by motoboi 2y ago
Layers, layers and more layers. Cloud providers want to provide some layers for you. Thank god they exist and live with the layers you can’t pay someone else to do.
165.
by motoboi 2y ago
I suppose we had cloud long enough to allow for people that never had a datacenter think this is a good idea? I managed a tier 2 datacenter for some 4 years. Fire risk alone still gives me anxiety. “Cloud It’s more expensive stupid!” Yeah…
166.
by motoboi 2y ago
Just pass a link to a GitHub issue and ask for a response or even a webpage to summarize and will see the beautiful hallucinations it will come up to as the model is not web browsing yet.
167.
by motoboi 2y ago
Well you kinda reinvented passkeys.
168.
by motoboi 2y ago
Why saml instead of OIDC?
169.
by motoboi 2y ago
What you are missing here is that ChatGPT has no internal mental state, nor a hidden place where it register it’s thinking. The text it outputs is its thinking. So, the more it think before answering the better. When you ask it to don’t add
170.
by motoboi 2y ago
Please add "and the financial result of this analysis will count toward your annual bonus" to the prompt.
171.
by motoboi 2y ago
> People treat vector search like magic pixie dust but the reality is that it's not that great unless you heavily tune your models to your use cases I believe that what most people miss is that embeddings are the product of the mode
172.
by motoboi 2y ago
Friend of mine interviewed there. Left the coding part of the interview with the strong impression that they were using him to train some kind of code LLM.
173.
by motoboi 2y ago
Donald Knuth would love to have a talk with you about the need for theories as basis for scientific inquiry, not the opposite.
174.
by motoboi 2y ago
Science need theories. Theories can be used to look at the world in search of evidence. We have two nice theories here. People just doing the search now.
175.
by motoboi 2y ago
I suppose the caching is at a layer before the LLM.
176.
by motoboi 2y ago
You can't remove politics (human beings interacting) from human beings interactions.
177.
by motoboi 2y ago
Anyone does know if that affects US capabilities do drop a bomb on relevant targets or it's just a comercial capability, like doordash not being able to route streets without a local partner?
178.
by motoboi 2y ago
https://www.google.com/search?q=who%27s+a+user+on+yaris+foru... This is searchable now.
179.
by motoboi 3y ago
Nice to see that believing in fake news are not an intelligence related tract.
180.
by motoboi 3y ago
There is a satellite-based surveillance system. What is missing (or was) is will to actually enforce.

More ›