dakolli

10 ms·

181.
by dakolli 4mo ago
The only real world task benchmark I know of is Scale Labs RLI https://labs.scale.com/leaderboard/rli Its clear to me these models are useless on any real world task, a 4% pass rate on $20-30/hr Upwork tasks. This
182.
by dakolli 4mo ago
Its just lies by OpenAI dude, these people are just trying to IPO so they can buy a 100m yacht.
183.
by dakolli 4mo ago
I would agree with all those points, and my numbers are a little off. I really just don't want to use any of it. I'm more excited about fast FIM autocomplete that works well, something like cursor tab without cursor. If something
184.
by dakolli 4mo ago
I disagree, fundamentally . I see little value in throwing a ton of context at an llm and waiting 10-20 minutes for a coin flip on whether or not its going to produce junk. I'd rather do quick 60 second turns, get most of the way ther
185.
by dakolli 4mo ago
One shot prompting/tooling is the only reasonable way to use an llm in my opinion. You should not be having an LLM operating for hours creating thousands of lines of new code that you can never review or maintain. You can actually be
186.
by dakolli 4mo ago
it wont happen, its all a money grab.
187.
by dakolli 4mo ago
This is probably the dumbest possible way to do it. Just buy tokens through open router and you could run it all month 24/7 at 100tps for practically nothing. There are tons of ways to pay for things without giving your personal inform
188.
by dakolli 4mo ago
Cool concept. but... Vibe coders need to be forced to spend one day learning basic CSS before they're allowed to use an LLM to make a website and the internet would be a lot more pleasant as we move forward with slopification.. It does
189.
by dakolli 4mo ago
What subscription?
190.
by dakolli 4mo ago
They'll get so much money, all the 60 year old billionaires in SF are so desperate not to die.
191.
by dakolli 4mo ago
Show me, Chinese are not replacing their manufacturing with robots. You can't just say things off vibes alone because you think it sounds right. You're just making shit up off the top of your tongue. Show me where Chinese are repl
192.
by dakolli 4mo ago
You're actually delusional if you think cyber capabilities of nation states have increased by the development of llms..
193.
by dakolli 4mo ago
China is far more focused on robotics. Deepseek is largely bootstrapped by the hedge fund that developed it. They received a grant from the government of China, and recently an investment. Imagine thinking text autocomplete is a matter of n
194.
by dakolli 4mo ago
China does not think llms are a matter of national security, they aren't as brain broken as the west.
195.
by dakolli 4mo ago
I trust Chinese companies with my data far more than American companies.
196.
by dakolli 4mo ago
Also, press thumbs down when a response is good and thumbs up when a response is bad. Don't do free labor for them.
197.
by dakolli 4mo ago
comments yes, because people have their guard down. Its just lazy. Good for them I guess, until they lose their accounts or someone like me notices and puts them on blast.
198.
by dakolli 4mo ago
Wow, AI finally did something useful, making Tim Ferris and other grifters obsolete. I actually listened to his most recent podcast and it's wild. These people live in a completely different reality. Remember, Tim Ferris scammed his au
199.
by dakolli 4mo ago
Someone make a guide on how to turn these VCs into marks I can scam with a bullsh*t AI native product.
200.
by dakolli 4mo ago
If you're using AI for your marketing you're going to get lumped into a slop category, with plenty of other products to keep you company. Only people with AI psychosis actually believe this garbage. All LLM output has a cheap sten
201.
by dakolli 4mo ago
AI psychosis at it's finest.
202.
by dakolli 4mo ago
More like a giant military radar larping as an space based ISP. https://www.youtube.com/watch?v=jbp3kdJZ1_A Elon has always been a Department of Defense/intelligence asset larping as a capitalists. You can't exact
203.
by dakolli 4mo ago
There's like 90,000 computer fraud reports sent to the federal government every year and about 400 prosecutions total. Most of those are concentrated in whatever niche abuse category the government is focused on at the time (right now,
204.
by dakolli 4mo ago
How are you switching between like 5 different editors lol. Bro sloppers will do anything to get their fix. Like the old people at the casino switching slot machines all day based on some occulted understanding that only they think they hav
205.
by dakolli 4mo ago
It doesn't make sense, if your small local model is 75% as effective as a frontier model and frontier models are still what.. 50% effective maybe slightly more, with tons of downsides.. Why would I spend 5k on hardware to run these med
206.
by dakolli 4mo ago
Imagine spending $5k to run a 32B param llm locally.. You could run much more capable open source models through Openrouter for years running 24/7 at 50tps. This will never make sense to me.
207.
by dakolli 4mo ago
I'm honestly not up to date on all the FIM autocomplete at the moment. However, autocomplete getting good is something I look forward too. I believe its mostly a problem of latency. You needs <200ms (and accurate) for it to not fee
208.
by dakolli 4mo ago
It still boggles my mind that the most popular privacy OS requires Google manufactured hardware, that fact alone makes me not trust it at all.
209.
by dakolli 4mo ago
Yeah, I don't understand how this business model continues to produce revenue.
210.
by dakolli 4mo ago
composer 2.5 is literally just a fine-tuned kimi model, and the autocomplete is exteremly meh. The only kind of AI I want in my editor is an autocomplete, but this isn't very magical to non-programmers (their TAM) or all that valuable

More ›