sweca

Ahh I didn't realize it was completely free. That's pretty crazy

> We are making this change to focus our resources on enhancing the main Tenor experience.

Is it really that hard to just keep the API running?

SHA-256 should not be used as a KDF. Use Argon2 instead.

Jemini is such a nice touch

This is because when you click a shared TikTok link, your account and the sharer's accounts are associated in a social graph. The sharer will see your account as a suggested friend and vice versa.

Limitless Pendant gets support for another year with existing customers getting full access without the need to continue paying for a subscription.

Rewind is being sunset.

hahaha j'adore que le mot québécois "calisse" est inclus

The best feature is capturing photos and recording life's moments with a button, but that's also not something I'd want to introduce to my friend group for its own privacy and ethical reasons (vs. the implicit consent you have with the motions of pulling out a phone)

I have Claude Pro and peak hours are in the afternoon and at night for me in EST

Google Pixel marketing is doing wonders for Gemini in young populations. I have been seeing a lot more of their phones in my generation's hands.

Custom TPUs ftw

Honestly, the best part about Gemini, especially as a consumer product, is their super lax, or lack thereof, ratelimits. They never have capacity issues, unlike Claude which always feels slow or sometimes outright rejects requests during peak hours. Gemini is constantly speedy and has extremely generous context window limits on the Gemini apps.

Raycast[0] has Gemini support in their AI offering and it's native, fast and intuitive.

[0] https://raycast.com/ai

Haha, yea. When I was writing this blog post (we had already pushed to production) I came across that post when searching for sources to confirm my findings. Antonio and I were so surprised at the coincidence.

It just means a virtual onsite type interview process. Real time interviews for both skill and cultural fit.

My company actually encourages the use of AI. My interview process was one relatively complex take home, an explanation of my solutions and thinking, then a live "onsite" (via zoom) where I had to code in front of a senior engineer while thinking aloud.

If I was incompetent, I could've shoved the problem into o1 on ChatGPT and probably solved the problems, but I wouldn't have been able to provide insight into why I made the design choices I made and how I optimized my solutions, and that would've ultimately gotten me thrown out of the candidate pool.

DO sucks anyways. Hetzner is better.

Tailscale also has funnel which would be more secure for public exposure.

You're going to face a challenge of scalability. With so many results like Google, you can't fit every embeddings vector in memory. You'll need to work a database hybrid of in memory caching and disk persistent indices.

It's an SPA framework so that's pretty expected

Because it's novel and interesting. This person wrote a super optimized dependency free implementation of GPT in C.

the repository description sums it up already: Effortlessly convert Spotify links to your preferred streaming service

I want to see SRP authentication implemented in more apps and services, so I decided to make a modern client and server library for JS runtimes.

If you are on the pay as you go model your data is exempted from training.

> When you're using Paid Services, Google doesn't use your prompts (including associated system instructions, cached content, and files such as images, videos, or documents) or responses to improve our products, and will process your prompts and responses in accordance with the Data Processing Addendum for Products Where Google is a Data Processor. This data may be stored transiently or cached in any country in which Google or its agents maintain facilities.

https://ai.google.dev/gemini-api/terms

No Human eval benchmark result?

Ollama has added support in a prerelease version.

https://github.com/ollama/ollama/releases/tag/v0.3.10-rc1

This is a really heartwarming story

No, it's not a standard library, but my point is UUID is standard and thus present in almost every language with a one line abstraction.

True but the appeal for most developers is it's simple to implement. Virtually every language has a UUID library that works in one line or code.

Like in Go, it's just uuid.New().String() vs using crypto/rand to read random data, convert it into Base64 of hex... which will take more lines and effort.