cosmosgenius

5 ms·

1.
by cosmosgenius 6mo ago
I would use it if it gives marker for time-to-respond, time-to-hire etc metrics. Metrics not from the company but verified and from others.
2.
by cosmosgenius 9mo ago
Is this any better than just doing Hotspot with wifi bridge? I just have my hotspot on my pixel for my devices to connect to. Pixel itself is connected to whatever "public wifi" is there.
3.
by cosmosgenius 9mo ago
for postgres specifically pgbackrest works well. Using in a home doing backups to r2 and local s3.
4.
by cosmosgenius 9mo ago
Android phones are not inherently linked to a google account, atleast not in the same way an iPhone is.
5.
by cosmosgenius 10mo ago
I can second 3blue1brown as intro, but anything practical needs practise. Get your local 12th grade or undergrad math book which has problems to solve will bring more focus and faster learning. Have a list of problems to solve in your head
6.
by cosmosgenius 10mo ago
Browser handling is way more personal than any other piece of software. It need not be open source licensed but being able to compile and install it from source the exact binary (minus signing) is a huge plus is today's world. Otherwis
7.
by cosmosgenius 10mo ago
Quantum compute would definite make a leap to moving closer to AGI. Calculating probability vector is very natural for quantum computer or more precisely any analog compute system would do. qubits==size(vocab) with some acceptable precision
8.
by cosmosgenius 10mo ago
Maybe it is the algorithms. But just by doing a op for an 10^25 param llm is definitely not feasible on todays hardware. Emergent properties does happen at high density. Emergent properties might even look as AGI.
9.
by cosmosgenius 10mo ago
LMAO! That Bruce Almighty reference had me rolling. Good one.
10.
by cosmosgenius 10mo ago
AGI might be possible with more Param+Data scaling for LLM. It is not completely within the realm of impossible given that there is no proof yet of "limits" of LLM. Current limitation is definitely on the hardware side.
11.
by cosmosgenius 10mo ago
Starting with an isolated POC showing the vector being exploited would help. I like gooogle project zero mainly for this.
12.
by cosmosgenius 11mo ago
The condescension is coming from "We want to make sure we trully understand what you're struggling with." which is completely dismissing the points they made in the post. Basics "surely there is more to it than just what
13.
by cosmosgenius 11mo ago
one thing which causes problem with git for me is collaborative work without using "git server". This usually comes up at homelab situation with no access a "git server" or ssh server. One thing with jj is i can use exis
14.
by cosmosgenius 1y ago
EV Certs used to do exactly that for me, that is until browser stopped make the visuals of it special. Don't think it would be even viable today given the short expiry (which is a good thing) of TLS certs necessary for browser https:&
15.
by cosmosgenius 1y ago
Maybe AI doing rebase, code chunking and commits will help? This kinda makes sense. Reviewable+mergables chunks do make the code way faster to merge.
16.
by cosmosgenius 1y ago
Does anyone know a low memory disk based search solution for home usage? By low memory i mean in the range of 256MB maybe 1GB. I am almost convinced of mellisearch but still looking. Elasticsearch would have been perfect but the memory usag
17.
Agile Otter Blog: There Is No Automatic Reset for Engineering(agileotter.blogspot.com)
2 points by cosmosgenius 2y ago|0 comments
18.
by cosmosgenius 2y ago
By having a 5 min group meeting where only the manager speaks. Manager is not allowed to talk about what direction the team should go and what the team should do. Anything else will give instant insights on whats going on at the isolated to
19.
by cosmosgenius 2y ago
Kagi is not mentioned true, but there are always talks w.r.t using site, intitle, time (before 2018 eg) filter to improve search. That indirectly can be considered as "having low quality" from one POV.
20.
VictoriaLogs v1.0.0 Release(github.com)
2 points by cosmosgenius 2y ago|0 comments
21.
by cosmosgenius 2y ago
Their 12b nemo model is very good in a homelab compared llama models. This is for story creation.
22.
by cosmosgenius 2y ago
Prompt to Output quality vary by a large amount between models IMO. The equivalent analogy would "lets switch programming language for this solved problem".
23.
by cosmosgenius 2y ago
The moment I saw the screenshot I was like "oh now I know what is was missing"
24.
Traefik with SSL: Understanding Acme(imagineer.in)
1 points by cosmosgenius 2y ago|0 comments
25.
Traefik 3.0 GA Has Landed(traefik.io)
7 points by cosmosgenius 2y ago|0 comments
26.
by cosmosgenius 2y ago
(derived, reproduceable-ish) sounds like a security horror O_o.
27.
by cosmosgenius 2y ago
I wanted to use Passkeys from the initial spec stage. The UX seemed far more superior (the closest I think is passwordless via email). But the more I wanted to use Passkeys are more scary it got, basically the gut feeling of losing control.
28.
by cosmosgenius 3y ago
I chuckled when i saw your username. syslog producing "Oh god" for "vault condition".
29.
Django REST framework 3.15.0(django-rest-framework.org)
149 points by cosmosgenius 3y ago|67 comments
30.
by cosmosgenius 3y ago
I just started using them specifically to test docker container implementation (Correctness of Dockerfile, entrypoint etc.)

More ›