andy99

36 ms·

691.
by andy99 1y ago
I suspect that ad powered LLMs will also dramatically sacrifice quality- they are a cost center suddenly, no reason to run a model with 10s or 100s of billions of parameters when some 500m thing will provide a minimally plausible scaffold f
692.
by andy99 1y ago
I think the bigger issue is outsourcing all or most thought to a LLM (data labeler really). Using tools you don't understand from first principles has always been commonplace and not really an issue. But not thinking anymore is new for
693.
by andy99 1y ago
I don't know if this is the right place to comment but I agree with you and noticed a lot of people are equating "survivable" cold/hot with not needing tunnels. In a city, most people are working in office towers and not
694.
by andy99 1y ago
> Would you rather spend four months in stasis waiting This is a false dichotomy. Hiring slowly doesn't mean doing nothing. It's really more like "do you (CTO) want to slow down on building and become a manager now?"
695.
by andy99 1y ago
This is a serious challenge in relation to hiring also. If you want to pay for good talent, and so are prepared to pay good money, how do you avoid people who are there for the money.
696.
by andy99 1y ago
I don't think government should be involved here, but what they can do is (a) always provide alternatives where interacting with government doesn't require a smartphone or apps, and (b) mandate the same for regulated or essential
697.
by andy99 1y ago
I'd guess it worse for us because it's a vector for disease. (And grosser to see for related evolutionary reasons). There's probably a greater volume in heavily trafficked places vs similar predators. Otherwise doubtful tha
698.
by andy99 1y ago
> Mount Everest – known as Sagarmatha in Nepali – to national parks in Norway and Aotearoa – known as New Zealand to English speakers Very hard to read around this stuff
699.
Edmonton radio station's use of AI host part of trend(nationalpost.com)
3 points by andy99 1y ago|0 comments
700.
by andy99 1y ago
> I have noticed how the majority of programmers I worked with do not have a mental model of the code or what it describes Possibly the same idea: lots of people at work don't appear to think about what they are trying to achieve an
701.
by andy99 1y ago
This is the opposite of a world model and is very much how machine learning works. It's purely correlative without some underlying theory (the world model). The article's (absurd) example of a person not being a mushroom gives a s
702.
by andy99 1y ago
> But the reality is how can someone small protect their blog or content from AI training bots? A paywall. In reality, what some want is to get all the benefits of having their content on the open internet while still controlling who get
703.
by andy99 1y ago
Orwell wrote about using metaphors (of which delve is one) Never use a metaphor, simile, or other figure of speech which you are used to seeing in print. At this point it's irrelevant of you're using AI or not, these words
704.
by andy99 1y ago
illusion For who's benefit?
705.
by andy99 1y ago
If you believe the bitter lesson, all the handwavy "engineering" is better done with more data. Someone likely would have written the same thing as this 8 years ago about what it would take to get current LLM performance. So I don
706.
by andy99 1y ago
https://archive.ph/GvICi
707.
by andy99 1y ago
I would group that kind of pricing with dark patterns, and I tend not no but anything when I see that. It's like designing the free shipping threshold to be a couple of dollars short of common purchases, buy 1 get the second for $5, et
708.
by andy99 1y ago
Yes, the problem is insurance, not who pays for healthcare. There's not universal government funded car maintenance and it works just fine for the most part. Insuring predictable, recurring costs is a scam. There's room, either pr
709.
by andy99 1y ago
What is special about healthcare in this regard and where should be draw the line? This argument could be used to justify socializing all kinds of costs, almost anything up to UBI.
710.
by andy99 1y ago
> mRNABench Just curious, in other areas of ML, I think it's widely acknowledged that benchmarks have pretty limited real world value, just end up getting saturated, and (my view) are all pretty correlated, regardless of their osten
711.
by andy99 1y ago
Is there evidence of that? That seems to have been the prevailing view over the last many years, and it's not clear to me that it's improved anything. There seems to be more homeless camps, more petty crime, more drugs.
712.
by andy99 1y ago
Well e.g keyword search (or looking something up in the index of a book) does not necessarily retrieve what you want, it just matches words. There's nothing special about semantic similarity in this regard, certainly no parallel with v
713.
by andy99 1y ago
I always thought it was the third derivative of position
714.
by andy99 1y ago
Also, enterprise customers are way more work, the sales cycle is longer, the compliance and onboarding stuff takes more time. There's a very legitimate case for building that into one's pricing, and it's really the same reaso
715.
by andy99 1y ago
There was the whole thing about the app for relationship pdiscussions, "Tea" I think, being apparently vibe coded and lacking proper security, leading to a data breach. More generally, I think vibe coding has replaced lots of mock
716.
by andy99 1y ago
The music thing is, I think, a related phenomenon I'll call the demolition man effect. Most people now are so weak and nonconfrontational that someone willing to break social norms can just walk all over them. This is because nobody wa
717.
by andy99 1y ago
$40B seems very low. I wouldn't be surprised to find the annual corporate churn on innovation / transformation / IT/IM new initiatives whatever you call it, is way higher than that. There's some subset of corporate
718.
by andy99 1y ago
> People who see everything through a political lens want to make everything political, because that’s how they gain power. Right- power over others seems to be the primary motivation for bringing up modern politics. I think it would pro
719.
by andy99 1y ago
I'd pay $800 if I thought it gave me something concrete. If it's just to try and jump through more hoops, it's like clipping coupons, it's not worth it and I'd rather just buy the thing I want. So it's weird th
720.
by andy99 1y ago
A couple things I've seen go by that make the connection. I haven't looked at them closely enough to have an opinion. > https://arxiv.org/abs/2507.12856 > https://justinchiu.netlify.app/

More ›