ttul

35 ms·

211.
by ttul 10mo ago
So I'm not the only guy doing this...
212.
by ttul 11mo ago
This is super cool work. I’ve built some new sampling techniques for flow matching models that encourage the model to take a “second look” by rewinding sampling to a midpoint and then running the clock forward again. This worked really well
213.
by ttul 11mo ago
I read that the pre-training model behind Gemini 3 has 10T parameters. That does not mean that the model they’re serving each day has 10T parameters. The online model is likely distilled from 10T down to something smaller, but I have not ha
214.
by ttul 11mo ago
1) Reading papers. 2) Reading "Deep Learning: Foundations and Concepts". 3) Taking Jeremy Howard's Fast.ai course
215.
by ttul 11mo ago
I wrote this, and honestly when I read it, I also want to reach for the LLM.
216.
by ttul 11mo ago
Prompt understanding will only ever be as good as the language embeddings that are fed into the model’s input. Google’s hardware can host massive models that will never be run on your desktop GPU. By contrast, Flux and its kin have to make
217.
by ttul 11mo ago
This post raises genuine concerns about the integration of large language models into creative and technical work, and the author writes with evident passion about what they perceive as a threat to human autonomy and craft. BUT… the piece s
218.
by ttul 11mo ago
This is a cool result. Deep learning image models are trained on enormous amounts of data and the information recorded in their weights continues to astonish me. Over in the Stable Diffusion space, hobbyists (as opposed to professional rese
219.
by ttul 11mo ago
My favorite benchmark is to analyze a very long audio file recording of a management meeting and produce very good notes along with a transcript labeling all the speakers. 2.5 was decently good at generating the summary, but it was terrible
220.
by ttul 1y ago
I'm going to guess that the difference is that Tailscale lets your machines find each other within a managed flat virtual network where as Iroh lets your applications talk to each other without any regard to which machine anything is r
221.
by ttul 1y ago
Yes, the newer image and video editing models have an LLM bolted onto them. The rich embeddings from the LLM are fed into a diffusion transformer (DiT) alongside a tokenized version of the input image. These two streams “tell” the model wha
222.
by ttul 1y ago
My rough guess is that they set a few workflows combining analytical and ML-based image manipulations to generate the training set. For instance, you can get a long way by having a segmentation model identify and mask various objects and th
223.
by ttul 1y ago
Flux backbone is too rigid. Very difficult to fine tune. Qwen is where it’s at these days.
224.
by ttul 1y ago
Unless you fine tune it… the guys of Qwen are amazing.
225.
by ttul 1y ago
Image editing model training is fascinating. One method for training image editing models involves using a second model to apply the inverse of the change you want the model to learn. Typically, the task you’re asking the second model to pe
226.
by ttul 1y ago
That is perhaps why they are billion dollar companies and why my company is not very successful.
227.
by ttul 1y ago
My impression is that `us-east-1` has the worst reliability track record of any region. We've always run our stuff in `us-west-2` and there has never been an outage that took us down in that region. By contrast, a few things that we ha
228.
by ttul 1y ago
There is always that point you reach where someone has to get on a plane with their hardware token and fly to another data centre to reset the thing that maintains the thing that gives keys to the thing that makes the whole world go round.
229.
by ttul 1y ago
It depends on what you want the model to do for you. If you want the model to complete text, then you would provide the input text unmasked followed by a number of masked tokens that it's the model's job to fill in. Perhaps your g
230.
by ttul 1y ago
This is a great summary. If you think about it a bit, text is an expanded representation of concepts meant for display on a two-dimensional surface that can then be read back by human eyes; our brains convert the two-dimensional information
231.
by ttul 1y ago
My view of the PC dev era was through the lens of a kid growing up in the 1980s with a dad who programmed for a living. My dad was a big fan of the Borland IDEs starting with Turbo Pascal and then moving on to the world of C and C++ by the
232.
by ttul 1y ago
This is what Codex is doing. The LM has been trained to work well with the kinds of tools that a solid developer would use to navigate and search around a code repository and then to reason about what it finds. It’s also really competent at
233.
by ttul 1y ago
Some of the best crypto scam companies were born in Vancouver…
234.
by ttul 1y ago
I concur here and would like to add that I worry less about sprawl when I know I can ask the agent to rework things in future. Yes, it will at times implement the same thing twice in two different files. Later, I’ll ask it to abstract that
235.
by ttul 1y ago
There are lots of comments here about how Figure’s website uses cherry-picked videos that are slickly produced. But that’s precisely what they must be doing at this stage. Yes, there’s some real tech being developed at Figure and it’s amazi
236.
by ttul 1y ago
I didn’t know this. Thank you for sharing.
237.
by ttul 1y ago
It's a shame that Canada also decided to shut the door on international students. The point was to ease the housing crisis (understandable) but the knock-on effect is to de-fund universities and surprisingly also public schools, which
238.
by ttul 1y ago
This article is pretty low quality IMHO. Lots of conjecture about the value of LLMs. We are far too early in this hype cycle to know what will come of this new tech. There is, however, no doubt that incredible amounts of capital are flowing
239.
by ttul 1y ago
This makes intuitive sense to me, because the system you are modeling is wide open and you’re competing against others who have the same information. Achieving much more than 51% accuracy would be extraordinary. But if you get 51% consisten
240.
by ttul 1y ago
"Wow, that argument with your ex-wife sure was spicy. And - hope I'm not prying in too much here - your lawyer may not be the best for this type of issue. May I suggest the firm Dewie Cheaterman & Howe? Y'know, I've

More ›