WhitneyLand

7 ms·

91.
by WhitneyLand 5mo ago
Wow, this is rough. Gemini Cli was already losing and it’s now being replaced by something they’re saying doesn’t yet have feature parity. Doesn’t seem likely to inspire defections from competitors. One could argue coding is only a use ca
92.
by WhitneyLand 5mo ago
So I made a mistake reading the article? So what? The point is you made two brigade style comments about my posts sounding suspiciously like an LLM and having hallucinations. Neither turned out to be true and I think a better response woul
93.
by WhitneyLand 5mo ago
Please see reply to your other comment on this thread.
94.
by WhitneyLand 5mo ago
So, you’re wrong on two counts. 1. Evidently you’re no longer able to distinguish AI from people as the whole comment was written by a human off the cuff. 2. The numbers are not hallucinations. It’s word on the street reporting, so yes it’s
95.
by WhitneyLand 5mo ago
I wrote this 100% off the top of my head on my phone while eating a sandwich. Ffs. edit: removed cursing you out. Sorry but this is frustrating. I don’t leave AI generated comments here (or anywhere else).
96.
by WhitneyLand 5mo ago
Their rationale might be that it’s size and intelligence are growing relative to the market. Fwiw it’s beating Claude Sonnet in most benchmarking (benchmaxxing?), yet they’ve priced it almost half off on a per token basis. Question is are y
97.
by WhitneyLand 5mo ago
Say what you want about Cursor but they don’t lack for ambition. Forking VS Code, going big on bleeding edge features like cloud agents, and now they’ve thrown down the gauntlet directly challenging frontier labs by training their own model
98.
by WhitneyLand 5mo ago
Could this task be a nice benchmark for computer use models? Would interesting to see the success rate for Claude Cowork or Codex’s equivalent feature.
99.
by WhitneyLand 5mo ago
How so? I’m not saying most of work doesn’t go into creating the drafting model or enabling a new head on the primary model, but the point is that however cool it is the result is, more weights. Speculative decoding requires code to be awa
100.
by WhitneyLand 5mo ago
Yeah important conceptually to remember MTP is kind of just more weights, but speculative decoding is the runtime algorithm that’s a significant add to whatever code is serving the model.
101.
by WhitneyLand 5mo ago
“This beats the latest Sonnet while running locally” Not really. - The benchmarks are based on F8_E4M3 and you’re not running that on any Mac. - Sonnet has a 1M token context window. This is 256k but again you’re probably not even getting t
102.
by WhitneyLand 5mo ago
Maybe <400ms is an inflection point but it sure isn’t optimal. “Productivity soars when a computer and its users interact at a pace (<400ms)”
103.
by WhitneyLand 6mo ago
The most important thing is keeping up the momentum to formalize more proofs and continue to strengthen the libraries and foundational work. If that momentum is strongest with Lean so be it. At the same time things become more machine verif
104.
by WhitneyLand 6mo ago
Did they not address how adaptive thinking has played in to all of this?
105.
by WhitneyLand 6mo ago
In case you wonder where the current trends come from. “Peter Thiel and Marc Andreessen have parlayed their extensive ties with the president into an unabashed assault on universities and institutional science. In private text messages leak
106.
by WhitneyLand 6mo ago
What does “revenue positive” even mean? It doesn’t mean profitable, it doesn’t mean cash flow positive. Are you just trying to say their revenue is greater than zero?
107.
by WhitneyLand 6mo ago
No, there is research in that direction and it shows some promise but that’s not what’s happening here.
108.
by WhitneyLand 6mo ago
>>For some reason Zed limits the Gemini 3.1 context to 200k tokens It’s not just Zed, CoPilot also reduces the capabilities and options available when using models directly. No thanks, definitely agree with the Open Router approach or
109.
by WhitneyLand 6mo ago
It’s not a flame war, and you’re not just sharing your experience and encouraging others to try it out. You’re making a claim, and I’m pointing out that it’s unsubstantiated and not consistent with any other source of data, including that i
110.
by WhitneyLand 6mo ago
“GLM5…better than Opus, Codex, Gemini…” What wild claim to make. Unsupported by benchmarks, unsupported by the consensus of the community, no evidence provided. Sounds like in another comment here even the GLM5 team concedes they are behind
111.
by WhitneyLand 6mo ago
Great music. Bright White Lightning, but didn’t see a track name. Overall, wow.
112.
by WhitneyLand 6mo ago
Wouldn’t this be way more expensive? Example 2TB: Google $10/mo vs S3 ~&45/mo? You could get cheaper that Google Drive with glacier tiers but that’s a different level of restrictions and still has retrieval fees.
113.
by WhitneyLand 6mo ago
Bullshit. This does not represent what real people are listening to, there are ways to game the system. The idea is explained by Rick Beato here: https://youtu.be/rGremoYVMPc
114.
by WhitneyLand 6mo ago
The features here don’t seem game changing. The most compelling parts are mostly already available in Claude or Codex or their related apps and services. The biggest concern is that if you want to use SOTA models I don’t see how they can
115.
by WhitneyLand 6mo ago
Saying his name like “girdle”, is the closest English pronunciation I’ve seen. The actual German ö is hard for me to figure out without having a native speaker around to practice with.
116.
by WhitneyLand 6mo ago
StepFun is an interesting model. If you haven’t heard of it yet there’s some good discussion here: https://news.ycombinator.com/item?id=47069179
117.
by WhitneyLand 6mo ago
A bit misleading to say they take 14x less memory, no one is doing inference with 16-bit models.
118.
by WhitneyLand 6mo ago
Call me a killjoy I hate April fools jokes.
119.
by WhitneyLand 6mo ago
Great quote from Hilbert, I think it’s also a useful thought for software development. “The edifice of science is not raised like a dwelling, in which the foundations are first firmly laid and only then one proceeds to construct and to enla
120.
by WhitneyLand 7mo ago
“The output is no longer 1-bit” I thought that constraint was the whole idea?

More ›