swalsh

10 ms·

151.
by swalsh 2y ago
Yeah, there are some categories with choices, but 9/10 use cases the .net ecosystem is Trader Joes.
152.
by swalsh 2y ago
SOTA AI companies are modern airlines. They're selling a capital intensive commodity that has little meaningful differentiation once you pass a certain threshold. Customers just want the lowest price available.
153.
by swalsh 2y ago
I suspect businesses will create VDI's or VM's for this express purpose. One because it scales better, and 2 because you can control what it has access to easier and isolate those functions.
154.
by swalsh 2y ago
I think Rabbit had the business model wrong though, I don't think automating UI's to order pizza is anywhere near as valuable as automating the app workflows for B2B users.
155.
by swalsh 2y ago
Building an entirely new world for agents to compute in is far more difficult than building an agent that can operate in a human world. However i'm sure over time people will start building bridges to make it easier/cheaper for a
156.
by swalsh 2y ago
I suspect these models have been getting smaller on the back-end, and the GPU's have been getting bigger. It's probably not a huge deal.
157.
Entropix: Entropy Based Sampling and Parallel Cot Decoding(github.com)
5 points by swalsh 2y ago|0 comments
158.
by swalsh 2y ago
The problem with today's technology is it is indistinguishable from magic. Sometimes the magic is real, sometimes it's an illusion. It's nearly impossible as a regular consumer not deeply knowledgable of the current capabil
159.
by swalsh 2y ago
My experience is more with Databricks, and their workflow system... but the concept is exactly the same. It let's data scientists work in the environment they work best in, and it makes it easier to productionize work. If you seperate
160.
by swalsh 2y ago
"You just can't compete with cheap Chinese labor" You sure can, we have cheap Mexican labor... and we have a much healthier trade relationship with Mexico.
161.
by swalsh 2y ago
I understand, but I was commenting specifically around the comment "US Manufacturing is reinvigorating" which it is, and which is more general than chips. I think when the Trump admin renegotiated trade policies (and I'll giv
162.
by swalsh 2y ago
There is a lag between policy and results, I don't think you can fairly say that because it's happening NOW it's biden's doing (though i'll give credit to Biden for now ending the changes Trump enacted). But Trump
163.
by swalsh 2y ago
"US manufacturing is about to be reinvigorated" Oh i'm going to be downvoted into oblivion for this one.... but I think this is a win we can give to Trump. It was a hard focus of his, I think he put the right people in place
164.
by swalsh 2y ago
But it also makes us not dependent on a place that China has their literal sights focused on.
165.
by swalsh 2y ago
We're not going to give you training data... for a better user experience.
166.
by swalsh 2y ago
Models don't really predict the next word, they predict the next token. Strawberry is made up of multiple tokens, and the model doesn't truely understand the characters in it... so it tends to struggle.
167.
by swalsh 2y ago
Thanks for helping me move, enjoy your free crime!
168.
by swalsh 2y ago
Yeah, the opportunist managers are attracted to fresh value like sharks to blood. They always destroy it because they're trying to add in something that's good for their business (great story for upper management) but is either n
169.
by swalsh 2y ago
I'd argue decline lags value fundumentals. Sears decline took decades. The early metric is probably "viewer hours" per customer or some use base metric. That will probably dip before the revenue dips.
170.
by swalsh 2y ago
There's a few Steve Jobs quotes which have tanked the tech industry. The quote "If I'd ask customers what they wanted, they would've told me a faster horse." for example is absolutely horrible advice for 90% of com
171.
by swalsh 2y ago
I think there's a property which is hard to describe. But i've seen the opposite property countless times at companies. People, and leaders who are building "cool demos". Things that tell a great story, but somehow ne
172.
by swalsh 2y ago
Has the company not already declined? I've been subscribed for years, but I cant remember the last time I watched something on it. The next time I do a round of canceling subscriptions, there is a very real chance it's gone. I
173.
by swalsh 2y ago
I'd argue what they REALLY need is domain experts who can write a good prompt.
174.
by swalsh 2y ago
Common issue when new tech comes out. The people who know the tech, but not their companies business focus on the tech. Many of them will get promotions, and make their way up in the company. The company will lose likely millions, and th
175.
by swalsh 2y ago
Site has a database error, so can't read the report. But here's what it probably says. "80% of AI projects don't solve a critical customer need, and find themselves with low usage/sales, and eventually run out of
176.
by swalsh 2y ago
I LOVE LM studio, it's super convenient for testing model capabilities, and the OpenAI server makes it really easy to spin up a server and test. My typical process is to load it up in LM studio, test it, and when I'm happy with t
177.
Llama3 Just Got Ears(homebrew.ltd)
5 points by swalsh 2y ago|0 comments
178.
by swalsh 2y ago
I'm wary of stuff like this, probably really useful to rapidly iterate.... but what a maintence nightmare after 10 years and your schema has evolved 100 times, but you have existing customers in various state of completeness. I avoide
179.
by swalsh 2y ago
This would explain why the answers to some of the hardest issues i've worked on seem to come to me after I wake up.
180.
by swalsh 2y ago
I think googles peering agreements are possibly the only reason YouTube is viable as a free service. Hard to compete against a company who basically doesn't have to pay for bandwidth.

More ›