maeil

16 ms·

151.
by maeil 2y ago
Would it sell more than the additional sales that were obtained by the work done during the time saved through generating it with AI instead of finding an artist?
152.
by maeil 2y ago
Agreed that Sonnet still feels like the best all-round model. The new ones are at least on par with it for pure coding, or exceed it (r1, o1 both do IME) but don't generalize as well, especially to tasks with subjective answers. I find
153.
by maeil 2y ago
This agrees with my limited testing so far, but in a different way: o3 being better at coding and objective tasks, with the most recent Flash 2.0-thinking stronger at subjective tasks. Similarly, o3 seems better at shorter output sizes, bu
154.
by maeil 2y ago
It changes the landscape with its multifaceted approach .
155.
by maeil 2y ago
Have you seen Azure VM SKU naming? It's.. impressive.
156.
by maeil 2y ago
Azure seems to be the least performant of all.
157.
by maeil 2y ago
Sounds like it was ineffective then, despite an export ban being easier to uphold than what would amount to an import ban?
158.
by maeil 2y ago
What I haven't seen mentioned here yet - there's also a (smaller) cohort of people who genuinely believed the lies that Musk peddled regarding upcoming self-driving functionality and are unhappy about having been scammed. I anecdo
159.
by maeil 2y ago
It's expected that German brands are popular but that even Volvo gains so much ground is interesting. Is it due to the zero import tax between EU countries? Are those Volvos still even made in Sweden (or EU)? In the US, it seems like H
160.
by maeil 2y ago
Your country's biggest SW company is SAP, world infamous for their German column names, haha. Pretty sure it's the most widely used product in the world with non-English internals that people actually interact with - I'm sure
161.
by maeil 2y ago
Were these implementations already easily open source accessible at the time, with tens of thousands of people already actively using them on their computers? No, right? Doesn't seem feasible this time around.
162.
by maeil 2y ago
Oh you're 100% right! All evidence points toward them not doing so! My (potentially hot) take is that if they'd gottten spanked with 40% the first time around, or Biden had lost, there'd have been much more external pressure
163.
by maeil 2y ago
And here are 3 things I learned.
164.
by maeil 2y ago
Some of them are close, most are cheaper than Sonnet 3.5, all of them are multiples cheaper than O1.
165.
by maeil 2y ago
You're talking about them as separate groups, but it's likely that the majority of Americans who consider themselves part of "the left" have been voting for these piss-poor Dem candidates that have indeed willfully hav
166.
by maeil 2y ago
> I've been shifting more and more of my coding from Sonnet to DeepSeek V3 in recent weeks. For what purpose, considering Sonnet 3.5 still outperforms V3 on your own benchmarks (which also tracks with my personal experience comparin
167.
by maeil 2y ago
There was a large amount of this kind of content ("what is censored") when ChatGPT suddenly got mass popularity as well.
168.
by maeil 2y ago
Luckily they didn't build it for under $6M. Not even V3 let alone R1.
169.
by maeil 2y ago
What are you talking about? Are you aware that the DeepSeek API has gone down for over a day and DeepSeek claims it's due to a large-scale cyberattack? [1] HN post about it got 20 points and 5 comments, completely ignored.[2] This one
170.
by maeil 2y ago
As far as Ive seen there's already at least 4 alternative providers out there and that's within days of release. Then there's all the distillations and all the projects to rebuild it. It's not trivial to run it but als
171.
by maeil 2y ago
Gemini 2.0-flash-thinking shows reasoning steps, it's only OpenAI who don't. Anthropic gives their system prompts. Looks to be SamA wanting to keep his company shrouded in mystery to appear stronger than it is.
172.
by maeil 2y ago
IME OpenAI doesn't block Hong Kong for one, while Anthropic does. As an aside, this made Cloudflare's AI Gateway an unusable product, as they hilariously put a node in HK so you never know when your Anthropic request is randomly g
173.
by maeil 2y ago
Agreed with your rant, but this time they won't get it both ways. The cats not getting back in the bag, the papers are out there, the knowledge is out there. Let's say the US bans DeepSeek from operating in the country. Firstly, j
174.
by maeil 2y ago
Don't worry, no such result will happen, as the other party is already showing they have just as little interest in actually changing things - they're going to keep everything "same old". [1] [1] https://www.n
175.
by maeil 2y ago
Don't worry, OpenAI definitely didn't improperly obtain any of their training data nudge nudge wink wink .
176.
by maeil 2y ago
To be fair, I recently posted a prolonged Anthropic complete API outage (when Sonnet 3.5 was best in town) and it got zero traction, so it's not unimaginable. Agreed it's a bit suprising though.
177.
by maeil 2y ago
Doing calculations like these is a really bad idea because they muddy the waters, putting the focus on disputable numbers - unless you're willing to do a proper study and really dive into it. It's extremely tiring seeing all these
178.
by maeil 2y ago
Medieval? Par for the course for modern dictators, see Putin or Xi Jinping. It's the default modus operandi.
179.
by maeil 2y ago
Would WaPo have stopped endorsing if it was set to endorse Bezos' favoured candidate? Of those that stopped endorsing, what's the ratio by candidate they would've endorsed? If only the outlets who were set to endorse candidat
180.
by maeil 2y ago
I wouldn't be surprised if LinkedIn is eventually excluded from training datasets. It's inclusion trains the LLMs to generate slop, which the providers will eventually be looking to reduce.

More ›