IanCal

17 ms·

331.
by IanCal 8mo ago
I’m always wary of anything that has such a clear example of a case that LLMs “don’t do” yet is trivially achieved by saying “review”. The slack message result for example saying they’re the lead designer but nothing else (with clearer and
332.
by IanCal 8mo ago
You really must be able to understand the difference between liking a thing and being addicted to a thing ? If not it’s probably worth just starting with basic definitions of addiction.
333.
by IanCal 8mo ago
The HF page suggests yes, with vllm. > We've worked hand-in-hand with the vLLM team to have production-grade support for Voxtral Mini 4B Realtime 2602 with vLLM. Special thanks goes out to Joshua Deng, Yu Luo, Chen Zhang, Nick Hill,
334.
by IanCal 8mo ago
If you use something like youtube-dlp you can download the audio from the meetings, and you could try things out in mistrals ai studio. You could use their api (they have this snippet): ```curl -X POST " https://api.mistral.a
335.
by IanCal 8mo ago
If your neighbours have so many kids that the newborn poop smell is filling your home if you open the window for a while and making your curtains and clothes stink of it because it’s that strong you are well within your rights to complain
336.
by IanCal 8mo ago
> You can’t instruct individual agents to contribute to global knowledge Of course you can. What makes you think you can’t? They can use task trackers like anyone. Also you can have review on prs, stages where you ask for review and then
337.
by IanCal 8mo ago
If you smoke you might not realise just how absolutely awful it smells and how it sticks to material. Sounds of children don’t.
338.
by IanCal 8mo ago
These aren’t tests against autonomous cars though these are tests against what would happen if you used, say, gpt4o to figure out what to do.
339.
by IanCal 8mo ago
Overkill I’m sure for many things but I’m curious as to whether there’s a TLA kind of solution for this sort of thing. It feels like it could although it depends how well modelled things are (also aware this is a 30s thought and lots of bet
340.
by IanCal 8mo ago
> Without a strong propsition as to why I, as a customer, would use an agent over not doing so? Whats the value prop? Same reason you’d use an automated approach to anything - you want to solve a problem and want to solve it along some P
341.
by IanCal 9mo ago
Yes largely, although they’ve trained a model specifically for this task rather than using the base model and a bit of prompting.
342.
by IanCal 9mo ago
I'm not sure whether there are parts of this done for claude but those other ones are layers on top of the usual LLMs we see. This seems to be a bit different, in that there's a different model trained specifically for splitting u
343.
by IanCal 9mo ago
Probably not, though additions maybe - I added the feature where the sculpt tool turns as you move it around if I recall right, many moons ago - I don’t think it was that hard but was a useful change.
344.
by IanCal 9mo ago
I think you massively underestimate the number of useful apps that are crud and a bit of business logic and styling. They’re useful, can genuinely take time to build, can be unique every time, and yet not brand new research projects.
345.
by IanCal 9mo ago
> if I now throw yet another (pointless) popup for you each time you install an app, are you OK with it? It would be an extremely minor issue, definitely not rise to the level of having multiple phones being easier. It’d be a few button
346.
by IanCal 9mo ago
Because it’s way easier to completely fuck up a system with running arbitrary commands on it while in use than it is by changing your code. It’s a massive step up in power and a massive drop in how much you can scrutinise a change (to zero)
347.
by IanCal 9mo ago
I’ll share another great version of Beowulf- Bea Wolf. Based on kids, with fantastic artwork and a great story/version. My kids absolutely love me reading this and I absolutely love reading it as a large passed down story of battles.
348.
by IanCal 9mo ago
I would absolutely put ssh access to the prod server way above submitting a pr for danger, that’s an enormous step up in permissions.
349.
by IanCal 9mo ago
A few points: 1. I think you have mixed up assistance and expertise. They talk about not needing a human in the loop for verification and to continue search but not about initial starts. Those are quite different. One well specified task
350.
by IanCal 9mo ago
Do you think Walmart should be handing your credit card numbers out? Genetic profiles of you? Is there any limit or do you think if you walk into a space whoever owns that space can get and do whatever they want with any information you mig
351.
by IanCal 9mo ago
Th consent is about tracking and your data, not specifically cookies. If you accept them tracking and selling your data then deleting cookies only impacts one way that happens.
352.
by IanCal 9mo ago
> What does it do when the model wants to return something else, You can build that into your structure, same as you would for allowing error values to be returned from a system.
353.
by IanCal 9mo ago
One part I like about LLMs is that they can smooth over the rough edges in programming. Lots of people can build pretty complicated spreadsheets, can break down a problem into clear discrete tasks, or can at least look at a set of steps and
354.
by IanCal 9mo ago
They've got a huge amount of space, solar has a low cost and provides an additional consumer to build out yet more capacity for supplying the world. > Wouldn't it be better to just go with nuclear If this is legit : https:&#x2
355.
by IanCal 9mo ago
That doesn’t matter to the point, which is stored history misses the way in which things moved from state A to state B.
356.
by IanCal 9mo ago
> Insurance that is maximally responsive to patient health changes in terms of cost (ie making healthier people pay less) ends up being an inefficient way of just having people pay for their healthcare directly. That's true for pred
357.
by IanCal 9mo ago
We could use markdown.
358.
by IanCal 9mo ago
Ideally IMO this would be accept headers. You're asking for the same semantic content but a different format . I'm not sure if there's a nice way of specifying html but in a minimal sense (we do quality with images, perhap
359.
by IanCal 9mo ago
It is, the content loads first, then the js for the cookie banner, then the favicon. If the js fails to load (I blocked the request as a test) the page loads just fine, it isn't blocked by that.
360.
by IanCal 10mo ago
I’ve often wanted a setup where it became physically harder to send an email to me the more unread ones I have to deal with. Like having to cram an extra letter into a pigeon hole that’s already full.

More ›