CTO at Kalvium and KalviumLabs. Ex-google, hackerrank. Focussed on improving Higher Education with democratic access to great tutoring pathways. AI pragmatist.
Submissions
Judging Technology – By Ran Prieur
ranprieur.com
1 points·by anilgulecha··1 comments
Internal AI policy at Zerodha: “We will not fire anyone..”
twitter.com
2 points·by anilgulecha··1 comments
Dimorphos – asteroid satellite with upcoming planned impact
en.wikipedia.org
2 points·by anilgulecha··0 comments
Attacking the Performance of Machine Learning Systems
schneier.com
1 points·by anilgulecha··0 comments
The SQLite OS Interface or “VFS”
sqlite.org
3 points·by anilgulecha··0 comments
REPL, an Effective Tool for Developers
medium.com
2 points·by anilgulecha··0 comments
Facebook Plans on Backdooring WhatsApp?
schneier.com
361 points·by anilgulecha··107 comments
Show HN: Bit-Rate Challenge: Think 320k mp3 files are better than 96k?
gist.github.com
2 points·by anilgulecha··0 comments
“More on my LinkedIn Account” – Bruce Schneier on his fake account
schneier.com
2 points·by anilgulecha··0 comments
CrossCheck: Google Partnering with First Draft and Newsrooms
blog.google
1 points·by anilgulecha··0 comments
Lectures on Digital Photography
sites.google.com
7 points·by anilgulecha··0 comments
“I Peeked into My Node_Modules Directory, You Won’t Believe What Happened Next”
medium.com
4 points·by anilgulecha··0 comments
Ask HN: Can we have a 'bookmark' function for HN submissions and comments?
9 points·by anilgulecha··10 comments
Ask HN: Uses for old touchscreen devices?
3 points·by anilgulecha··1 comments
A website to log you out of the mot popular websites
superlogout.com
3 points·by anilgulecha··1 comments
Ask HN: Anyone use derbyjs?
1 points·by anilgulecha··0 comments
Substance editor's funding bid to build an open publishing platform
I built a harness in the cloud with slack like interface for this, allowing multiple people to join. It allows chat members to tag action items, work on files and sync to GitHub. My team uses it to coordinate on our task and teM context together, setup meetings which are auto transcripted and ingested into the attached wiki, and connect to calendar to setup meetings. Built it for our needs.
Didn't really consider putting it out in public. Is there a viable product out of this? How much would you pay.
Camera app should negate the need. most pictures are of people and scenary, and 99.99% of the time the intent is to take the photo in the right order.
Simple totally offline ONNX models exist, whcih should make it trivial to categorize the right orientation. Acceleometer/magnetometer can feed this, but should not be the default.
Just do this and avoid the hassle of rotating at all!
I understand the frustration (I'm an avid fdroid user across many many devices). But this article comes off as childish with the virus/trojan/"malware vendor".
With such an article, many (including perhaps google) get the ammo to disregard what fdroid says, by branding them as childish/not to be taken seriously. for eg: no reputable news org is going to post this.
We can get creative. quick ideas: Send it by printed post. pass it around people to people. an email needs to be added in with some process, and can only get one TXT update value a week.
Many ways of adding friction to obtaining the updatable value - which a human owning a domain would be happy to do, but a squatter would not want to.
Hence the "in the first year or two". Some more human effort to showcase proof early on, then the domain is solidified for you like iwth any other registrar. This is something like captcha/bcrypt - a single instance isn't a burden, but doing it at scale is costly.
> You cannot transfer ownership killing any value you added
I think this is by design. The domain should be for personal use - hence free.
The "one free domain per person" isn't the interesting part really - that will be hard to police unless domain name is a function of ID proof (avoids squatting).
0) The actual intersting part of a new TLD can be growing reputation by post-facto taking away a domain without recourse in case of squatting. Instead of adversarial takedowns (which produce false positives as noted), let anyone challenge an inactive domain in the first year or two.
1) If they can figure out a mechanism for moving a domain from "assigned" -> "squatted".
2) Domain must match (or derive from) a verified identity - e.g. your domain is a hash/slug of your government ID. Makes squatting structurally impossible because you can't claim someone else's name / gov (Sign in with passkeys linked to a national ID).
3) Proof of human effort, reduced with time - require periodic renewal with proof-of-use (DNS TXt updates, through a flow hard to automate).
4) Kill speculative market - domains are non-sellable and non-transferable - always go back to the free pool, and stay there for 30 days mandatorily.
Some mix of these could be the right structure for a trule high-reputation, free domain.
The author should call out size of binaries (compiled and generated), the interpreter as well since that's teh selling point. if it's tiny from memory requiremetns , shoudl also call out the same, potentially benchmarking against similar (luajit?).
It's not technically a translation, it's a re-implementation, with test suites acting as the destination. If it was a file by file translation your argument would have been valid.
> We believe it would be good for the world to have the option to slow or temporarily pause frontier AI development to enable societal structures and alignment research to keep up with the advance of the technology. The Anthropic Institute will conduct research—in collaboration with many others—and take actions to help build the systems that a credible slowdown or pause would require.
Interesting - they're commiting to kickoff policy conventions to organize a world-slowdown of frontier LLM building. If they actually are able to crack it, this will give a much needed breather IMO. As exciting as the last ~6 months have been, there's some bigger questions to go answer now.
Disagree is such a loose/wimpy study. Add in a grounded/expected response, and then it becomes a better benchmark (because it'll force the author to actually think about choices presented to the LLM).
1) That LLM/Agents are being pushed and not adopted. I see plenty of deep adoption by junior folks.
2) The unit economics don't work out. From the details on every model so far - each model is wildly profitable over it's amotized time-frame. It's just that money is used upfront for the next model, and each next model is significantly more costly to train. The best case argument instead is - this will not last and we'll pour more on some models, than see in it's revenue.
I think realistically these form the core of the thesis, and IMO, and hence it's conclusions are a bit off the mark.
I have anosmia, triggered by AERD/polyps. I have been mostly without the sense for the past ~12 years, but int eh past year have had bouts of smell again, via a doc who finally diagnosed AERD, and suggested steroid intervention + mepolizumab.
CTO at Kalvium and KalviumLabs. Ex-google, hackerrank. Focussed on improving Higher Education with democratic access to great tutoring pathways. AI pragmatist.