ttul

39 ms·

571.
by ttul 3y ago
Yes. I discussed this issue with the author of the ComfyUI IP-Adapter nodes. It would doubtless be handy if someone could end-to-end train a higher resolution IP-Adapter model that integrated its own variant of CLIPVision that is not subjec
572.
by ttul 3y ago
Hi Alex Amazing work. I scanned the paper and dusted off my aging memories of Jeremy Howard’s course. Will your model live happily alongside the existing SD infrastructure such as ControlNet, IPAdapter, and the like? Obviously we will have
573.
by ttul 3y ago
Yes, because why would anyone have a child when they can’t be sure of finding housing or childcare?
574.
by ttul 3y ago
BC changed its primary care billing model dramatically in 2023 to address the problem. The new model pays doctors based on a combination of their patient load and the complexity of their cases. The typical family doctor pay rises as a resul
575.
by ttul 3y ago
Well, in BC, you can go here [1], enter anyone’s name, and find their complete criminal justice history in the province. For $6, you can get copies of all the filings, such as the alleged charges. If their charges were not pursued further,
576.
by ttul 3y ago
What they’re referring to is the court records that show whether someone was ever put through any kind of criminal justice process. If someone is arrested and charged, that record is publicly accessible to anyone for a few dollars. If the c
577.
by ttul 3y ago
Frankly, this is yet another article justifying why you must never speak to the police.
578.
by ttul 3y ago
On a positive note, I am seeing more sites offering Passkeys, which integrate natively with browsers and provide vastly superior security compared with passwords. We won’t ever eliminate the human error that leads to password theft. At leas
579.
by ttul 3y ago
This company is reasonably well run. After going back to AWS, they doubled their revenue and things are going well. They are not incompetent. They did earnestly try to cut their costs and just didn’t see the iceberg.
580.
by ttul 3y ago
Nothing a few million can’t make better with some effective altruism sprinkled about.
581.
by ttul 3y ago
Indeed. This is the hangover from the era of cheap money and private equity gone wild. It’s a healthy adjustment for the industry that puts great people back in circulation, to be picked up by healthier companies that didn’t blow their brai
582.
by ttul 3y ago
Incompetence. Take my friend’s company for instance. They were frustrated paying $60K/mo to Amazon so their brilliant sysadmin bought $600K of servers and moved them into a cheap colo. Over Christmas, everything died, and the brilliant
583.
by ttul 3y ago
Future?
584.
by ttul 3y ago
It’s an interesting example of a new technology that is inferior to the old one because the old one is superior in important ways (speed, ease of manipulation, and established toolsets).
585.
by ttul 3y ago
It’s the “we will make this so easy for you that you never want to switch” moat. Definitely akin to Slack, which also has the integration glue to keep you on their platform. Even though there are many Slack alternatives now that are really
586.
by ttul 3y ago
I don’t agree with you here. OpenAI should be free to train on your data assuming you agreed to that in the terms of service (and yes, you did). If they ask for a little more money in exchange for not having that valuable information in tra
587.
by ttul 3y ago
I don’t think so. Here are some differences: 1. OpenAI is building in a brand new space. Mobile phones are well established. LLMs as a product for consumers are brand new. While it’s true Rabbit are trying to merge LLMs with mobile, the ele
588.
by ttul 3y ago
I’ve spent a few hours reading the Mamba paper, examining the code, and watching Yannick’s YouTube dive into the paper. If you took discrete time signal processing in school, you will have what you need to work through the math. Here is my
589.
by ttul 3y ago
That seems unlikely given that both Apple and Google have been employing the world's top ML scientists for years, have unlimited budgets, have better access to customers than any startup (to find out what customers want), and ... need
590.
by ttul 3y ago
d9 = dooooooooo
591.
by ttul 3y ago
My experience - anecdotal - is that it is more consistently speedy as well.
592.
by ttul 3y ago
I was surprised to see in the paper that they appear to ask for an answer to a multiple-choice question, and only after getting the answer to the multiple-choice question do they ask the model for an explanation. The model will do much bett
593.
by ttul 3y ago
I hung out with Theo and his team years ago at a security conference. At the time, Subversion was the new hotness, sweeping away CVS as the source code management system de jour. I asked Theo why they didn’t already move to Subversion. His
594.
by ttul 3y ago
And here is the layout: https://www.dropbox.com/scl/fi/ksxru0lmnp0oha356fdby/layout.... With apologies, it was a CRC-8. Still so many nodes and so many wires…
595.
by ttul 3y ago
Simulated annealing has been useful for decades. My undergrad VLSI project was to lay out a 32-bit CRC chip and test it with SPICE (an analog simulation tool). To get the capacitance down, you need short wires. But a CRC-32 has many many in
596.
by ttul 3y ago
I believe Mozilla falls into a difficult non-profit grey area. Having another browser on the market isn’t as critical to humanity as providing anti-malarial medicine to millions of Africans, yet I think we can all agree that a world without
597.
by ttul 3y ago
I don’t know enough to answer your question, sadly.
598.
by ttul 3y ago
My IQ is orders of magnitude lower than the authors of the paper, but I did my best to work through it anyway. I studied CE and have the basic control theory background and undergrad level discrete time systems intuition. It would take much
599.
by ttul 3y ago
This is a great way to consume papers. If there’s one thing LLMs know, it’s machine learning literature!
600.
by ttul 3y ago
You are correct and thank you for the correction.

More ›