ignoramous

7 ms·

91.
by ignoramous 2mo ago
"The thing" isn't free [0]: It is cheap by US/EU income standards. > who is the customer If no one else, then definitely those that can only budget $1/mo to $5/mo. Probably 100s of millions, if not billion
92.
by ignoramous 2mo ago
Isn't Reasonix [0] DeepSeek's own harness? Or, are they building a new one? [0] https://reasonix.io/
93.
by ignoramous 2mo ago
DeepSeek v4 API rates are only matched by Xiaomi's MiMo v2.5 series. Quite similar in capabilities too (at least before today's update).
94.
by ignoramous 2mo ago
MiniMax's another lab that's known for relatively smaller models (their latest, M3 is 295b) that punch way above its weight.
95.
by ignoramous 2mo ago
Could OpenAI be running a different model for high + max ?
96.
by ignoramous 2mo ago
DeepSeek v4 Pro & MiMo v2.5 Pro (Opus 4.6 quality models for code) are insanely cheap for agent-driven work due to their super low cached-input prices ($0.0036/mtok) [0]. For Luna, the cached-input price drop isn't disclosed i
97.
by ignoramous 2mo ago
I was in a similar situation and probably sent a cringy, emotional email because when I read it back now after all these years, I am embarrassed at how desperate my tone was. It all worked out well, though; I learnt to handle rejections bet
98.
by ignoramous 2mo ago
> the next task is SO HARD that it doesn't make sense to not use frontier How does tokenless quantify "so hard"? > 1 turn to crack versus 100 turns for Deepseek to crack Interesting definition for a "frontier
99.
by ignoramous 2mo ago
Thanks! https://github.com/openai/codex-security/issues/52
100.
by ignoramous 2mo ago
Heh, reminds me of this talk at !!conf: Tail Call Optimization: The Musical , https://youtu.be/-PX0BV9hGZY (2019).
101.
by ignoramous 2mo ago
Exciting! Is there any open GitHub issue we can track?
102.
by ignoramous 2mo ago
Something about the Pine tree that the British also planted forests of them in the Levant: https://www.trtworld.com/article/9ab8ec0a341a
103.
by ignoramous 2mo ago
> why on earth are we still using simple numbers to describe which process to connect to? Why not use names instead. Not super popular, but: https://en.wikipedia.org/wiki/SRV_record
104.
by ignoramous 2mo ago
> Kurt personally working with me to help debug our deployments Typical Kurt in Founder Mode . For the first 3+ years (immediately following their "launch hn"), Kurt was almost ever present in the forums. > claim strong
105.
by ignoramous 2mo ago
Scott was instrumental in turning around Docker (which was "broken" [0]), after a $23m Series B raise from Arjun Sethi at Tribe Capital: https://medium.com/better-programming/the-greatest-startup-t... &#x2F
106.
by ignoramous 3mo ago
> spyware to be installed as a device administrator, granted accessibility privileges, and then given every possible Android permission via AppOps Correct, and apps using these setups (even for genuine reasons) have been under the cosh
107.
LoongFlow: Loop engineering and structured memory for long-horizon tasks(github.com)
2 points by ignoramous 3mo ago|0 comments
108.
by ignoramous 3mo ago
No.
109.
by ignoramous 3mo ago
The article overlooks security implications from spyware, which is a huge problem not only for financial applications, but personal safety, too [0]. Per FTC, a stalkerware will: geo locate, read call list & record calls, read notificati
110.
by ignoramous 3mo ago
You need update your blog posts. Our whole stack is radically open source — frontend and backend alike, Apache-2.0 licensed — and so is everything behind this benchmark. That is how a benchmark number earns trust: verifiability, not hyp
111.
by ignoramous 3mo ago
Do-able in many low-income countries but it'd be borderline miserable, by modern consumption standards.
112.
by ignoramous 3mo ago
Mixture of Models , perhaps? tbf to OP, the setup they're proposing has also been recently evangelized by other "ai gateway" products (like OpenRouter, JusCode, Fireworks etc), so there's likely something useful here.
113.
by ignoramous 3mo ago
I see eventuality here as job cuts or salary cuts. Don't think that day is far when "software people" are paid as if they were taxi drivers.
114.
by ignoramous 3mo ago
> I can't use it for theoretical physics because I can't evaluate the responses. A tool, even if it is a chisel, in the hands of a master sculptor would obviously result in a wildly different outcome.
115.
by ignoramous 3mo ago
The free OpenCode Zen models (I've only had the need for DeepSeek v4 Flash & MiMo v2.5 for one-time / throw-away tasks) did work, for me, with GitHub Copilot & Reasonix.
116.
by ignoramous 3mo ago
Sounds like 3 more iterations on K3/Fable, and we'll able to ask it to break computational hardness assumptions like DLP?
117.
by ignoramous 3mo ago
> I'm bad at math ... he includes the GPT5 prompts for his conversation, which are easier to follow You were kidding, right?
118.
by ignoramous 3mo ago
Simplest way is to signup to OpenRouter and filter out all non ZDR (zero data retention) providers. Paying "API rates" however can prove expensive compared to coding plans (for instance, MiniMax $20/mo coding plan allows 1.7b
119.
by ignoramous 3mo ago
> focusing on what they can get wins in like speed instead Speed as a differentiator has always been Google's thing. They (used to?) show the microseconds it took to query & rank web-scale search results. Chrome, notoriously
120.
by ignoramous 3mo ago
> "The highest tier Chinese models are not more economical than US frontier models. Try GLM 5.2 and see how much it costs to do real work. I did, and it was more expensive than GPT 5.6." This is a flatly false statement. It m

More ›