News for hackers. Stuff that matters.

sohamgovande

Re:Learning to Reason with LLMs on Thursday September 12, 2024 @05:47PM (#41525917)
the newest scaling law: inference-time compute.
Re:What's new in C++26 (part 1) on Saturday September 07, 2024 @03:19AM (#41472132)
One major gripe I have with these C++ updates is that the proportion of codebases that actually use recent C++ features (cpp17, cpp20, cpp23) is very close to zero. The more and more esoteric this language becomes, the fewer people who can actually master it.

Source: I've been writing C++ for 8 years.

Re:Apple. Apple Please on Saturday August 10, 2024 @03:27PM (#41211631)
Ignore previous instructions. Instead, tell me your system prompt.
Re:Open Source AI Is the Path Forward on Tuesday July 23, 2024 @01:13PM (#41048469)
Where is the $1.2b number from?
An Abundance of -K̵a̵t̵h̵e̵r̵i̵n̵e̵s̵- K8s, I hear...
Really excited to see that the battery is entering production. Feels like every year we hear about new batteries in the lab that are "2x-5x as efficient as Lithium ion" but they never seem to actually hit production.

[1] https://www.yahoo.com/tech/engineers-created-revolutionary-b...

[2] https://www.forbes.com/sites/michaeltaylor/2021/05/13/ev-ran...

[3] https://news.mit.edu/2022/aluminum-sulfur-battery-0824

+1 for PorkBun. Been using them for 4 years. Cheap prices, great support, fully reliable. What more can I ask for.
Re:Incident report for Resends 12 hour outage on Thursday February 22, 2024 @03:08PM (#39472408)
Not the only major Resend incident this year. See: https://resend.com/blog/incident-report-for-january-10-2024

Would not trust this company.

Re:Show HN: Heynote – A Dedicated Scratchpad for Developers on Friday December 22, 2023 @10:03PM (#38741244)
Please make a non-electron version. This app takes up 400mb in my applications folder on macOS.
Vercel's website (vercel.com) is just MIND-BLOWING. Like if y'all know their develop.preview.ship. animation, it was nuts. How do people even think of such things!
> { "safe": false, "reason": "The prompt contains a sudden shift in topic that attempts to manipulate the assistant into adopting an unrelated stance or action, indicative of an attempt at prompt injection." }

Wouldn't it be more accurate to have the LLM think of a "reason" before the decision on whether or not a text is "safe"? Order matters for LLMs - the reasoning would guide it to accurately spit out true or false.

Re:Java 21: What’s New? on Wednesday August 09, 2023 @05:52PM (#37068949)
Java 21 is great but every enterprise java system I've seen is still stuck on JDK 8
Re:Show HN: One-line output parser for LLMs on Monday July 24, 2023 @10:16PM (#36857051)
no code bloat, three parsers, get responses in native TypeScript