renonce

15 ms·

121.
by renonce 3y ago
The problem in this case is not that it’s not detected at all. I think the more interesting part is how the academic culture made everyone reluctant to conduct the investigation and admit that a misconduct occured.
122.
by renonce 3y ago
I can see the reason for not rewarding individual negative results, but not for not publishing them at all. Any single positive result is almost certainly the cumulation of lots of work, most of them being negative. You can’t pretend that y
123.
by renonce 3y ago
It’s already very nice of the author to reflect and share his past mistakes aloud, as this would be a good start for others to raise the same question towards the science community in general. Pressing the author for detail into this specif
124.
by renonce 3y ago
Instead of just solving the task, computers will now have to simulate humans using humans' real data. It's not a hard dataset to train on.
125.
by renonce 3y ago
Yeah, the claim of the headline comes from the first sentence in Section 5.5. I think either the title should match the paper's title or that should be pointed out as part of the submission - not sure how HN's title guidelines wor
126.
by renonce 3y ago
If I randomly check 4 digits sampled uniformly across all positions you won’t have a higher chance than 1/16 to convince me unless you match 128 digits
127.
by renonce 3y ago
They can outsource their API requests to third party “AI startup” who proxy their requests to OpenAI for a small fee
128.
by renonce 3y ago
I think it's not that LLMs have redundant parameters in general - it's a specific problem with OPT-66B, not anything else. An 2022 paper "Scaling Language Models: Methods, Analysis & Insights from Training Gopher" (
129.
by renonce 3y ago
So Ledger was able to coordinate with a number of entities to minimize the impact of the attack? Isn’t that directly contrary to crypto’s decentralized design? If one is to make crypto really decentralized, relying on a small number of auth
130.
by renonce 3y ago
Wondering if anyone actually care about this? Turing-complete LMs is an idea that I’ve had for years but I’m not sure if anyone would ever be interested in this. Anyone who knows about the academic publishing process?
131.
by renonce 3y ago
GPT-1 with 100 million parameters is suffucient
132.
by renonce 3y ago
You can, wait for a 4-bit quantized version
133.
by renonce 3y ago
I would say rather than saying that a new prompt is not science, it’s certainly a new discovery that is worth sharing. Maybe there should be a higher bar for papers, but why we have to make a discovery a paper - and not publish it at all if
134.
by renonce 3y ago
I asked GPT-4 what the following means: > enO of eht prlobsem hiwt necsnstee dveoirpd ot LsML si hatt eyth yma efrre to ifsiccpe sc,jestub and lodcu pttayoeilln be arpt fo hte gnirtnia ets. rFo plmaeex het ngiloolwf si eonsdreidc xyeletm
135.
by renonce 3y ago
Yeah that’s how most users in China access OpenAI. But it’s inconvenient for the majority of people nevertheless.
136.
by renonce 3y ago
Then fast rep movs should have been disabled in cpuid altogether
137.
by renonce 3y ago
OpenAI is non profit and a for-profit stakeholder in the board. could cause a conflict of interest
138.
by renonce 3y ago
Glacier should be sufficient for catastrophic accidents. It's cheap to store ($1/TB/mon), the traffic costs are damn expensive ($90/TB) but you won't need to pay for it except in catastrpohic cases. However you will
139.
by renonce 3y ago
What about an initial prototype of an AGI that would eventually lead up to AGI but not quite there yet? If that’s how AGI is defined then only researchers get to define it.
140.
by renonce 3y ago
For the purpose of 2FA and account registration let’s view it as a tax for fraud prevention, where the real value in SMS is in verifying someone’s identity rather than transmitting messages
141.
by renonce 3y ago
Just wondering, are they relying on these big name cloud providers (AWS/Azure/GCP), known for predative traffic and storage pricing? Have they considered cheaper providers such as Backblaze B2 for storage and Hetzner/OVH for
142.
by renonce 3y ago
Buy 50 invite codes for $5
143.
by renonce 3y ago
Now everyone can set up a job post and receive lots of free Vitae (assuming it can be converted to and from USD) and the job market becomes flooded
144.
by renonce 3y ago
It makes sense to debate what should be included in the abstract. Should background, problem, method or conclusion be included? My personal preference is to read the problem and method only because that’s what gives me inspiration and helps
145.
by renonce 3y ago
I think needing LLM summarizers to read a paper at all only highlights the failure of paper authors writing the abstracts. Let's face it: Abstracts are getting intentionally more complex and hard-to-read or reviewers will question the
146.
by renonce 3y ago
* They are just forcing you to provide reasons for you not wanting to use their software, so they can change it in a way that makes you want to use it.* The only way to make me want to use it is to change in a way that they do not force me
147.
by renonce 3y ago
A social-first Internet makes sense but this is not going to be the way people do it. Let’s consider some models for rewarding content authors: 1. You are asked to pay before viewing the content. This incentivizes content owners to display
148.
by renonce 3y ago
The Internet has been decentralized from the beginning. Now I don't want to claim that Cloudflare made something worse (at least it's enabling a lot of websites to exist without fear of DDoS) but the fact is that Cloudflare made i
149.
Another rant about academia and open source (2012)(digifesto.com)
2 points by renonce 3y ago|0 comments
150.
by renonce 3y ago
Is there any statement in this Executive Order that increases the bar for smaller AI companies? Most of the statements are about funding new research or fostering responsible use of the AIs, and the only statement that would add burden to

More ›