memothon

8 ms·

1.
by memothon 3mo ago
Was this announced beforehand? How do you double prices for customers so abruptly with no transition period?
2.
by memothon 6mo ago
I'm always skeptical because you can make it pass the benchmarks, then you use it and it is not practically useful unlike an extremely general model. Cool work though, really excited for the potential of slimming down models.
3.
Show HN: Gcrunner – Run GitHub Actions on Google Cloud VMs(github.com)
1 points by memothon 6mo ago|0 comments
4.
by memothon 6mo ago
Yeah I agree with this. I will try to use it in earnest on my next project. That metric is the key piece. I don't know the right way to build an automated metric for a lot of the systems I want to build that will stand the test of time
5.
by memothon 6mo ago
I think the real problem with using DSPy is that many of the problems people are trying to solve with LLMs (agents, chat) don't have an obvious path to evaluate. You have to really think carefully on how to build up a training and eval
6.
by memothon 6mo ago
I've used browser rendering at work and it's quite nice. Most solutions in the crawling space are kind of scummy and designed for side-stepping robots.txt and not being a good citizen. A crawl endpoint is a very necessary addition
7.
by memothon 6mo ago
This model is probably really strange to people who haven't spent a ton of time working in something like Clojure.
8.
by memothon 8mo ago
Turso offers a cloud-hosted sqlite product. The idea is you can easily spin up per-tenant databases and have a "serverless" sqlite interface. It feels like it has a lot of the same downsides of hosted databases so I'm not sur
9.
by memothon 11mo ago
If you implement different local payment providers that you could see higher conversion rates.
10.
by memothon 1y ago
Lots of companies have automations with Zapier etc. to upload things like invoices or other documents directly to notion. Or someone gets emailed a document with an exploit and they upload it.
11.
by memothon 1y ago
Some people have startup credits
12.
by memothon 1y ago
I made a quick site so you can see what tools are using the most context and help control it, totally free and in your browser. https://claude-code-analysis.pages.dev/
13.
Show HN: Memothon – a custom GPT with spaced-repetion review(memothon.com)
1 points by memothon 2y ago|0 comments
14.
by memothon 2y ago
Yeah this was a huge foot gun
15.
by memothon 2y ago
Those WAL entries streamed to object storage I think are just for backups. Each DO is globally unique (there's one DO with a given id running anywhere) and runs sqlite on its own local storage in that datacenter.
16.
by memothon 2y ago
I'm imagining your poor 90 year old aunt playing this wild game of Simon says with you and having no idea what's going on. Maybe just ask the cousin not to send any more money?
17.
by memothon 2y ago
Yeah I get what you're saying. totally agree.
18.
by memothon 2y ago
Replying to bookmark
19.
by memothon 2y ago
Good question, remains to be seen. Will need to see how it plays out!
20.
by memothon 2y ago
I don't disagree it is unsustainable, I'm really trying to be more precise about whether anyone is getting value out of the tools. I'm just really skeptical that nobody is getting value.
21.
by memothon 2y ago
Sorry I should have been more specific. The article does mention that OpenAI has huge revenue. > While The Information reported that OpenAI's revenue is $3.5 to $4.5 billion a year in July, The New York Times reported last week that
22.
by memothon 2y ago
I always find it really strange when articles like this claim nobody is paying for generative AI. I can't find reliable stats on this but there are at least a million ChatGPT Plus subscribers. Does that not count? It takes time for ne
23.
by memothon 2y ago
The A-Z of Product Manipulation would be a more apt title.
24.
by memothon 2y ago
UAE maybe?
25.
by memothon 2y ago
I think the discussion around "exponentials" with top-end LLM (think 3.5 sonnet, gpt-4 not the smaller models) scaling is really pointless. The heuristic we have for what to expect from performance is just scaling, which has worke
26.
by memothon 2y ago
In the specific case of the Drake track, hateful is an appropriate word. He was using Tupac's voice on the TaylorMade freestyle in a really disrespectful way that would borderline on hateful of his artistic legacy. Just read these lyri
27.
by memothon 2y ago
I think your post is taking these AI voices out of their original context. First let's consider CD -> streaming as a media change. Streaming didn't really exist when Tupac was around. But nobody would say putting Tupac's c
28.
by memothon 2y ago
Thanks all!
29.
by memothon 2y ago
Did you happen to remember the domain-specific LoRA startup?
30.
by memothon 2y ago
I really like the point about getting AI to ask you questions. The focus in the AI tutor world is basically a chatbot to ask questions of. But if you're trying to learn something, it's really helpful to have targeted questions ask

More ›