Filter
Exclude
Time range
-
Near
Replying to @kenpaquette
Butts . . . JT's Pal...Serial Labeller with Evan Solomon.
14
I just saw a bsky account that was just a normal person but they made themself into a labeller so they could have a square pfp holy aura
2
23
488
huggingface.co/datasets/Infa… 841 hours (310 videos total) at 256x144 res at 15 fps of terraria gameplay to do VPT style world model training. you just have to get your IDM (labeller) model first. was gonna do world models but got distracted so figured i should at least but up a cleaned dataset :)
1
9
979
This is the drum I’ve been beating for the past year - stealing enterprise data through contractors the way Mercor, buying the codebases of dead startups the way Handshake and others are - is cope for the fact that nobody has designed the right business model to incentivize companies like Ramp (that have some of the best people for a given domain, who use models day-to-day and know their failures inside-out) to partner with a third-party that can translate this feedback into RL envs. Such a partnership is needed because paying people full-time to generate RL envs creates envs that don’t reflect the frontier, because these creators grow disconnected from the real world. And because it’s more economically productive for now for SWE’d and other domain experts to actually work in their own domain full-time vs data labeling. E.g the best lawyers will earn more from being a lawyer than a data labeller. This is why I call this feedback ‘exhaust’ - it’s a by-product of doing the real work that is very economically valuable. Designing a partnership model where one resists the urge to pitch their investors that they’ll replace the other party is key. This is both a technical breakthrough (automating the above process) but also a business model and culture breakthrough. This is what I want to do for the next decade of my life because it’s the only way to access the highest quality RL data and to avoid adverse selection in data quality. If this sounds interesting, I’d like others to bounce ideas off of too.
Today we’re releasing Ramp SWE-Bench: a private, production-grounded coding benchmark created from real engineering problems we've faced at Ramp.
1
4
577
From the senses, data from the brain I believe is interpretted by mind which in turn has already been labelled, all the permutations and combinations that could ever be by the Great Labeller aka God Who is Knowledge Itself- what do you think of that?
1
4
Where the heck is my wire labeller… always something
1
40
Did you know? The 2200 Series Print and Apply labeller ensures every product leaves your line coded, readable and traceable. Built-in checks. Built for reliability. pyrotec.co.za/product/2200/ #PrintAndApply #PackMark #Traceability #Production #Coding
3
Replying to @lesandfurious
biseksüel çekim zaten queer bir çekim, yaptıkları bifobiklikten başka bir şey değil hani özel olarak tanımlamak istiyorsan da sapphic ve acchillean ile belirtebilirsin. sınırları gayet belli iki yönelimi alıp kendine göre dönüştüremezsin, labeller insanların kafalarına göre eğip bükmesi için yok ki KAPSADIKLARI insanları ifade etmek için var. yani sana uymayan bir label varsa onu rahat bırakmak bu kadar zor olmamalı, hayır her şey olamazsın yani çocuk oyunu yaptılar iyice
2
1
16
659
Replying to @GergelyOrosz
how a data labeller leads a AI research team at a MAG7 company is baffling, always knew meta was going to fall behind after the fiasco
4
476
After some thinking and prototyping, the only thing I'd tolerate annotating that data myself, is dump individual frames/head-crops, and then range-select them and classify into quadrants. This is fast but coarse. Since images are dumps from videos, there's continuity and I could average out like 5img/sec or so. I did it twice: once for front/left/right/back. And once again for front-left/front-right/back-left/back-right. Then we can marge this to get 8-bin orientation. Turns out @PINTO03091, the probably most GOATED labeller in history, converged to the same schema, which is what reminded me of my work yesterday.
1
1
40
2,856
Civilisations don't decline because people get weaker. They decline because the structures that produced good outcomes from self-interested people are dismantled, and the dismantling is rational for the people doing it. That is the Madisonian standard. It is the analytical anchor of everything I write. What works is what works. The label tells you about the labeller, not the labelled.
3
112
the fourth job application i made on telus is "ai content labeler intermediate (ai community)" the expected payout range for this gig is $6-$12/hr as a content labeller, you really don't need a specialized technical degree or previous ai experience to start your job is to help ai "see" and "understand" the world by sorting and tagging information. essentially, you are the teacher providing basic lessons that make ai smart and safe what you'll do: • content sorting: read short pieces of text and put them into categories so the ai learns how to organize information • simple tagging: look at images or text and add "tags" (like labels) to describe what is happening • basic fact-checking: verify that the names, dates, or simple facts provided by the ai match the information found in real-world sources • instruction following: check if the ai followed simple rules provided in a "playbook" or set of guidelines note that you'll go through a few subject-specific qualification exam that will determine your suitability for this position but don't worry, the exam is easier than mercor's and outlier's
quick update here, guys a real world understanding, generalist role that pays $34-$40/hr mercor is seeking sharp, analytically minded generalists to help train AI systems on real-world reasoning and visual understanding tasks • what you'll do?? you'll evaluate AI performance on ambiguous, multi-modal challenges that require common sense, spatial reasoning, and the ability to interpret real-world scenarios that don't fit neatly into a single domain • qualification?? the ideal candidate is a recent graduate from a selective university with strong general reasoning skills and intellectual curiosity someone who's comfortable with ambiguity, can think critically about visual and real-world tasks, and bring the kind of flexible problem-solving that comes from a rigorous academic background this is a flexible, remote, hourly role in which you'll tackle fascinating AI reasoning challenges on your own schedule apply here asap : t.mercor.com/WzTbN
1
4
250
This is series where I review tools listed on Viral Bucket. Today I am reviewing a Chrome extension I use daily. Build by @yuki_eliot X account labeller. Highlight your favorite person with your favourite color. PS. Also if you have already highlighted my name what color have you used?
5
1
12
516
Labelling someone extreme far right automatically renders the labeller extreme far left and a r*tard.
1
1
15
242
I am old. Dymo labeller
1
15