keeda

50 ms·

721.
by keeda 11mo ago
> I doubt it. I don't. This is simply the "drug dealer" model where the first hit is free. They know that once people are addicted, they will keep coming back. The question of course is, will they keep coming back? I thi
722.
by keeda 11mo ago
On the flip side, there are many indications that a huge part of corporate AI usage is "Shadow AI", i.e. where workers use GenAI without their employer's blessing or knowledge. E.g. https://cacm.acm.org/news&#
723.
by keeda 11mo ago
This goes back to how we define "novel problems." Is a dev building a typical CRUD webapp for some bespoke business purpose a "novel problem" or not? Reimplementing a well-known standard in a different language and infra
724.
by keeda 11mo ago
Unless everybody is writing the same code to solve the same exact problems over and over again, by definition LLMs are solving novel problems every time somebody prompts them for code. Sure, the fundamental algorithms and data structures an
725.
by keeda 11mo ago
Ehh, I dunno, it was really, really popular back then. I would bet that a non-trivial number of apps were built using VB by actual software engineers. Couldn't find concrete numbers but this article claims at its peak, 2/3rds of a
726.
by keeda 11mo ago
It's just a matter of economics. Where the costs of bugs in production is low, which is probably the vast majority of the software out there, extensive test coverage simply doesn't make economic sense. Something breaks in some nic
727.
by keeda 11mo ago
We all knew this was coming, but I had an experience a year+ ago that showed this ChatGPT already had this capability back then. It also showed how AI could truly differentiate itself from other ad vectors, potentially in an actually heplfu
728.
by keeda 11mo ago
Not sure about market studies, but I don't think the data is going to be very encouraging for this case either. E.g. in early 2024 Amazon introduced ads in Prime Video, which people were already paying for. I couldn't find more re
729.
by keeda 11mo ago
I dunno... Consider: 1. Token prices keep plummeting even as models are getting stronger. 2. Most models are being offered for free at a significant loss, so reducing costs would be critical to maintain some path to sustainability. 3. Every
730.
by keeda 11mo ago
It's not just Satya, the CEOs of all the hyperscalers are consistent in their messaging over the last several quarters that they are backlogged on capacity. Not only are they saying it, they are putting their money where their mouths a
731.
by keeda 11mo ago
> And just when was the last time you saw CEO of a company as big as Microsoft "caught lying to shareholders" about anything actually face any punishment? Actually curious, when was the last time we saw a CEO of a company as
732.
by keeda 11mo ago
I think it would help if either side could be more quantitative about their claims, and the problem is both narratives are usually rather weaselly. Let's take this section: >Deep learning thus far is shallow and has limited capacit
733.
by keeda 11mo ago
Hah so, I have a couple of old Mac Minis in my homelab that I use very occassionally but would like to access over the Internet (say, via Tailscale etc.) Ideally they'd be sleeping most of the time to conserve power, and only be kicked
734.
The State of Generative AI Adoption in 2025(stlouisfed.org)
2 points by keeda 11mo ago|1 comments
735.
by keeda 11mo ago
Study from the Federal Reserve Bank of St. Louis based on national survey data and other statistics indicating GenAI may already be impacting labor productivity at a national level. Relevant quotes: > When we feed these estimates into a
736.
by keeda 11mo ago
> Of course I still bill the time it would have taken me but for a professional it can be a great improvement. This may be a flippant comment, but it actually represents one of the reasons it is difficult to track GenAI usage and impac
737.
by keeda 11mo ago
Maybe the archive link stripped it out, but it would be really useful to look at the actual sources, because TFA seems to be, uhh, "selective" in what stats it presents. For example, this source (Alex Bick at the St Louis Federal
738.
by keeda 11mo ago
I meant it more as an observation than an optimistic prediction, really :-) The article is sound, but it's focus on large public failures disregards the vast, vast, vast majority of the universe of software projects that nobody really
739.
by keeda 11mo ago
But my point wasn't about AI helping or not, my point was AI will simply accelerate the natural trajectory of your organization. This is not a hypothetical, this is based on reports using large-scale data like DORA and DX: https:&#x2F
740.
by keeda 11mo ago
I think there is a ton more nuance, but can still be explained by a simple observation, which TFA hints at: "It's the economics, stupid." Engineering is the intersection of applied sciences, economics and business. The econom
741.
by keeda 11mo ago
Not really, by most indications AI seems to be an amplifier more than anything else. If you have strong discipline and quality control processes it amplifies your throughput, but if you don't, it amplifies your problems. (E.g. see the
742.
by keeda 11mo ago
Licensing and the perceived rigor it signifies is irrelevant to whether something can be considered "professional engineering." Engineering exists at the intersection of applied science, business and economics. So most software pr
743.
by keeda 11mo ago
> Weirdly, Windows Phone was aimed at consumers at a time when they really could have leveraged integrations with products like Exchange and Office to stand out. This is because a completely under-appreciated apsect of the iPhone revol
744.
DeepSeek writes insecure code if prompt mentions topics restricted in China(crowdstrike.com)
5 points by keeda 11mo ago|1 comments
745.
by keeda 11mo ago
Seems to be an in-the-wild, inverse instance of "Emergent Misalignment" as decribed in this paper: https://arxiv.org/abs/2502.17424 (Previously discussed here: https://news.ycombinator.com/ite
746.
by keeda 11mo ago
> Either reality has a left bias, or our training data does. Or these models are truly able to reason and are simply arriving at sensible conclusions! I kid, I kid. We don't know if models can truly reason ;-) However, it would be
747.
by keeda 11mo ago
I was peripherally looking into this for a similar problem domain: https://en.wikipedia.org/wiki/Private_set_intersection Related to Zero Knowledge Proofs, the advantage is that phone numbers need never be shared in cl
748.
by keeda 11mo ago
A labor economist I follow explains that this is just a side-effect of the shutdown, but other sources of employment data show that things have not deteriorated too much from the months prior. There have been some general, very slightly con
749.
by keeda 1y ago
I should have been more precise with my terms, but there is a difference between "food" and the "food industry" indicated by the likes of Nestle. Yes, everybody needs food. No, nobody needs the ultraprocessed junk Nestle
750.
by keeda 1y ago
https://en.wikipedia.org/wiki/Total_addressable_market

More ›