cbg0

13 ms·

241.
by cbg0 6mo ago
> all other European countries said it wouldn’t This is false, France made a misstep and even fired the head of French military intelligence because of the failure to predict the invasion and Germany was also notably skeptical. Other co
242.
by cbg0 6mo ago
I'm aware of how fast an SSD is, but you're probably only accessing your home NAS from your phone or laptop which most likely won't have more than 1gbps, if that. If you're a power user you probably already have high spe
243.
by cbg0 6mo ago
I doubt there's any FOMO to it from people using it for professional coding for example, you just want to have access to the best model as long as the value proposition is there. If I spend less time reviewing the work done because it&
244.
by cbg0 6mo ago
How is 1G insufficient for running a NAS over the internet?
245.
by cbg0 6mo ago
What a time to be alive to see people put "only" in front of "1gbps". We've come very far.
246.
by cbg0 6mo ago
The open models are still far behind GPT 5.4 and Claude if you're using them for building software.
247.
by cbg0 6mo ago
You have no idea how much slop people push out on daily basis.
248.
by cbg0 7mo ago
Is this the norm in US courts, evaluating a human's performance against LLMs?
249.
by cbg0 7mo ago
> Sandra Scott, MD, CEO of the One Brooklyn Health, a small hospital facing tight margins, agreed with this line of thinking, according to Crain’s. Does this CEO of a small hospital realize that their hospital will take the legal respons
250.
by cbg0 7mo ago
When you're on the hunt for VC cash "numbers go up" is the main criteria.
251.
by cbg0 7mo ago
> The EU is fundamentally flawed. There are no checks and balances You're missing a [citation needed] on that.
252.
by cbg0 7mo ago
When you say "hinders innovation" do you mean "doesn't maximize shareholder value" ?
253.
by cbg0 7mo ago
The range claims depend on the size of the battery pack. The Denza has a larger pack than what is quoted in the article. Also, the Chinese CLTC range ratings are overly optimistic with 1000km CLTC being ~820km WLTP or ~700km EPA.
254.
by cbg0 7mo ago
> at some point why doesn’t the CTO / CEO / etc. say “I’m going to go have conversations with the workers to get their perspective?” Why would you when all of the reports you're getting from your managers are 5/5 star
255.
by cbg0 7mo ago
Scores higher than Opus 4.6 on their in-house benchmark? Sounds legit.
256.
by cbg0 7mo ago
Not the only indicator, for sure.
257.
by cbg0 7mo ago
I love how many of these comments have em dashes in them and how many are just outright trolling.
258.
by cbg0 7mo ago
Based on the SWE-Bench it seems like 5.4 mini high is ~= GPT 5.4 low in terms of accuracy and price but the latency for mini is considerably higher at 254 seconds vs 171 seconds for GPT5.4. Probably a good option to run at lower effort leve
259.
by cbg0 7mo ago
No true scotsman, right?
260.
by cbg0 7mo ago
Perhaps this paper might be more relevant with regards to multi-agent pipelines https://arxiv.org/html/2404.04834v4
261.
by cbg0 7mo ago
It depends on what life you've lived so far.
262.
by cbg0 7mo ago
Unless you're clearing up the context for each description or processing them in parallel with subagents your context window will grow for each short description added to it making you hit those hour limits.
263.
by cbg0 7mo ago
Those 100 npm packages won't load themselves.
264.
by cbg0 7mo ago
It seems like every prompt is sent over to gopeek.ai and that's a pretty big thing you forgot to mention.
265.
by cbg0 7mo ago
There's quite a few asterisks that need to be appended to "independent".
266.
by cbg0 7mo ago
You only have to be worried if you're doing something illegal, like the guys in the article. Misfiling something won't land you in jail, just some fines at the most. Intent matters quite a bit.
267.
by cbg0 7mo ago
Try this: Keep your CLAUDE.md as simple as possible, disable skills, and request Opus to start a subagent for each of the files and process at most 10 at a time (so you don't get rate limited) and give it the instructions in the skill
268.
by cbg0 7mo ago
xhigh effort is actually pretty terrible for 5.2/5.3/5.4 models. Stick to medium/high as it overthinks less.
269.
by cbg0 8mo ago
Elevated errors on Opus 4.6 https://status.claude.com/
270.
by cbg0 8mo ago
This is very dependent on what kind of work you're asking the agent to do. For software, I've had quite a bit of success providing detailed API specifications and asking an LLM to build a client library for that. You can leave it

More ›