seizethecheese

7 ms·

91.
by seizethecheese 3mo ago
Perhaps only a small minority want quality out of information product, while most want satisfying simplicity. This is definitely how the news plays out. I recently built an ai chart product that attempts to add more nuance by injecting extr
92.
by seizethecheese 3mo ago
Yes I double checked the quote is not in the article. HN is probably the best place on the internet for people actually reading the article, but this being the top comment here suggests that the majority of voters still do not read the arti
93.
by seizethecheese 3mo ago
Texas -> Texan Kansas - Kansan Honduras -> Honduran It should be Dallan?
94.
by seizethecheese 3mo ago
We also don't fully appreciate how much this shaped the American Baby Boomer psyche. My dad talks of regular Nuclear War drills in elementary school.
95.
by seizethecheese 3mo ago
The shorter useful life implies less risk though, since therefore the projections don’t have to play out for very long.
96.
by seizethecheese 3mo ago
I think the stock wipeout would itself trigger a recession. If the hyperscalers needed to wipe out most of their income on interest expense they’d lose a large amount of their market capitalization. This could drop the stock market a huge a
97.
by seizethecheese 3mo ago
It clearly will be a big problem when the lending stops, there’s just no evidence of it in the article.
98.
by seizethecheese 3mo ago
Sure, I’m just saying the article doesn’t have any evidence of the lending stopping.
99.
by seizethecheese 3mo ago
The article transparently documents that people are lending to them.
100.
by seizethecheese 3mo ago
> AI’s insatiable need for debt has so far been matched by investors’ appetite for it, but they may turn nauseous on the belly-busting volumes coming from tech giants. Headline doesn't really match the facts in the article. The arti
101.
by seizethecheese 3mo ago
I’m curious what other professions have analogous issues. Police who seed crime? (Surely.) Developers who insert bugs?
102.
by seizethecheese 3mo ago
Their router classified prompts into difficulty buckets. This obviously won't work. Consider a senior developer routing work based only on the task description. Clearly you need to dig a bit deeper into the task. Saying routers don&#x2
103.
by seizethecheese 3mo ago
You’re right about a lot of this but there’s no evidence of trade slowing down. US exports are actually accelerating https://fred.stlouisfed.org/series/IEAXGS
104.
by seizethecheese 3mo ago
Flirting with one 9 of reliability http://status.claude.com/
105.
by seizethecheese 3mo ago
Ah I didn’t see the tabs in the benchmark chart. Performance is worse on the two coding benchmarks and cheaper which makes sense. I did read the blog post and I’m not sure what you mean by your reply. I don’t think the blog post addresses m
106.
by seizethecheese 3mo ago
Super interesting approach. It's probably novel. I can say this because I've been working on something similar (while building a code version of http://pellmell.ai ). I'm skeptical though. In order to pick which mo
107.
by seizethecheese 3mo ago
I’m colorblind and the weird thing is that it’s somewhat unpredictable when I will not see things. It’s not so simple as “can’t read red on green”. Shading and lighting matter enormously. An orange frisbee is harder to find in grass then a
108.
by seizethecheese 3mo ago
Claude Opus 4.8 was not able to stump open weight models and Opus 5 still can't (in this case Kimi K3 and GLM 5.2): https://pellmell.ai/s/35c98b86f9aa93e4ca713079d96b20f4
109.
by seizethecheese 3mo ago
Yep! There are a lot of models like this: https://artificialanalysis.ai/models/capabilities/coding?cos...
110.
by seizethecheese 3mo ago
Yep! I found this interesting article after banging my head against a wall for a long time: https://epoch.ai/gradient-updates/gpqa-diamond-whats-left
111.
by seizethecheese 3mo ago
Isn't the relevant benchmark RouterBench? https://arxiv.org/html/2403.12031v2#S7
112.
by seizethecheese 3mo ago
This looks different, it's not a main agent delegating to subagents, it's a router.
113.
by seizethecheese 3mo ago
They're obviously not claiming this is a new model that is fable-like at 1/3. It's a router that saves money by only using Fable when necessary. I don't think you should be calling someone's post a "scam"
114.
by seizethecheese 3mo ago
Sometimes expensive models are cheaper on easier tasks because they use fewer tokens, too.
115.
by seizethecheese 3mo ago
You definitely still would, but you need to pay the full input cost twice, so the equation really depends on how much first message vs repeat message matter.
116.
by seizethecheese 3mo ago
Fusion is a totally different approach, though similar in the sense that it leverages different models. Fusion generates many replies then synthesizes. This adds a ton of latency and cost, so it's going to be better only for cases wher
117.
by seizethecheese 3mo ago
If you want to try a similar idea that you can chat with immediately (free tier uses open weight models), I'm working on this: http://pellmell.ai
118.
by seizethecheese 3mo ago
OpenRouter's model router for coding isn't really sophisticated. This is very different.
119.
by seizethecheese 3mo ago
The Right Stuff by Tom Wolfe has a lot of interesting stuff on this wrt the first astronauts (who were sourced from test pilots). The test pilots saw the astronauts as second class citizens, at first, because they weren't actually doin
120.
by seizethecheese 3mo ago
I've been working on a similar project and I found that it's easy to replicate Fable results if you use saturated benchmarks. In my project, I wasted a huge amount of time trying to improve GPQA Diamond results above ~93% range. I

More ›