tikkun

I'm puzzled. Did Noam get $500mm+ from this?

Character.ai could've been at $100mm+ ARR if they did a bit more of a monetization push based on my very rough estimates. If it was an acquisition I would've been imagining $3b+ price range.

Huge get by Google! (Side note, Gemini 1.5's new alpha release from this week is now at the top of the lmsys leaderboard and sentiment on twitter for it is that it's strong, maybe as strong as sonnet 3.5, so it'll continue to be an interesting race between meta, openai, anthropic, gdm.)

Edit: Okay perhaps it's - Noam keeps his C.ai stock, gets a big pay package from Google ($5mm-$15mm/year kinda range? not sure). Most of the value in C.ai remains, and he keeps his stock.

The highest order bit about job search being broken at the moment is that the supply of jobs has fallen. It was bad before, too, but that's the primary thing making it worse. The main solution is to spend more time in person networking: find companies you like, and attend any tech talks they host at their offices.

Way better, nice

> We are excited to introduce Flux

I'd suggest re-wording the blog post intro, it reads as if it was created by Fal.

Specific phrases to change:

> Announcing Flux

(from the title)

> We are excited to introduce Flux

> Flux comes in three powerful variations:

This section also comes across as if you created it

> We invite you to try Flux for yourself.

Reads as if you're the creator

A general maxim: When something is a big deal, your response should make a bigger deal out of it than the complaints. $10 gift cards will feel like an insult to those impacted.

(Oh, and as mentioned in the existing discussion, it seems like the cards aren't even successfully redeeming.)

Existing discussion: https://news.ycombinator.com/item?id=41058261

I'd suggest using Perplexity.

Correct

This is definitely worse than no gift card. Insulting. A general maxim: When something is a big deal, your response should make a bigger deal out of it than the complaints. $10 says "We don't think this matters." Now watch as everyone explains precisely why it does. PR 101 fail.

My personal experience says Claude 3.5 Sonnet.

When using a prompt that involves thinking first, all three get it correct.

"Count how many rs are in the word strawberry. First, list each letter and indicate whether it's an r and tally as you go, and then give a count at the end."

Llama 405b: correct

Mistral Large 2: correct

Claude 3.5 Sonnet: correct

Links to chat with models that released this week:

Large 2 - https://chat.mistral.ai/chat

Llama 3.1 405b - https://www.llama2.ai/

I just tested Mistral Large 2 and Llama 3.1 405b on 5 prompts from my Claude history.

I'd rank as:

1. Sonnet 3.5

2. Large 2 and Llama 405b (similar, no clear winner between the two)

If you're using Claude, stick with it.

My Claude wishlist:

1. Smarter (yes, it's the most intelligent, and yes, I wish it was far smarter still)

2. Longer context window (1M+)

3. Native audio input including tone understanding

4. Fewer refusals and less moralizing when refusing

5. Faster

6. More tokens in output

Pick the one where you admire and like the founders more, and/or where you’ve used and believe in the product most. If you’re not sure, spend more time with the founders before choosing.

People who own Bigscreen headsets as well as Vision Pro, how do they compare for you?

Yes. Also, how much can Nvidia revenue go up from here? There are only a small handful of companies that want more than 100k H100s - if there are ten, that’s $30b of GPUs, which Nvidia would sell in less than 6 months, and then who keeps buying?

Or are we expecting a few companies to be buying 1 million plus H100s each next year

Avoid moralizing: Focus on delivering factual information or straightforward advice without judgment.

No JADE: Responses should not justify, argue, defend, or explain beyond the clear answer or guidance requested.

Be specific: Use observable, concrete details that can be verified or measured.

Use plain language: Avoid adjectives and marketing terms.

4o feels worse than 4 for me.

Eugh. Has anyone compiled a list of companies that do this, so I can avoid them? If anyone knows of other companies training on customer data without an easy highly visible toggle opt out, please comment them below.

Reselling the API access!

What's the subreddit?

Looks great, but I've become skeptical of Google demos and as such will believe it only after I've used it :)

What type of small businesses? I think you'd need to specify a specific category, otherwise there's not enough in common with the needs. Car dealership operators aren't going to look at the same site as people who run restaurants, most of the challenges are domain specific. HN = tech domain, which is specific enough to be relevant to most tech business people.

Another idea: let me see what’s newest to streaming that has good reviews. And what’s about to come out.

I currently use the rotten tomatoes new to streaming page for this; but I wish it was better.

The exceptional speech to text on the chatgpt iOS app is the main reason I use it. It’s so good. (I mean when you hit the microphone button in the bottom right, not the headphones / back and forth voice chat. Ie STT but no TTS.)

I’m assuming it uses whisper v3 plus gpt for post processing the whisper transcript.

I hope that with iOS 18 siri’s transcription is as good.

ChatGPT’s transcription is good enough that I can trust that even for a long 2 minute input, it’ll be 99-100% right. Whereas Siri I can’t trust that even for a single sentence.

The really good ones tend to make it to the HN homepage. Or, Twitter, if you want to drink from the firehose.

Best ones to play with:

Midjourney

Suno AI (music)

Claude

ChatGPT

Perplexity (a better version of “ChatGPT with bing”)

Then some more that others seem to like but I haven’t used / haven’t gotten into: Character AI and Pika Labs

I suspect these videos would be higher quality than most offerings from universities or business schools (given that it's a fast moving field, and most of the true experts are practitioners, and so if they're teaching, they're not teaching actual classes, at most they're doing youtube videos or interviews):

- https://www.youtube.com/watch?v=jkrNMKz9pWU (Jeremy Howard Hacker's Guide to LLMs)

- https://www.youtube.com/watch?v=zjkBMFhNj_g (Karpathy Intro to LLMs)

What's the pricing, if you know? I see a request pricing form, it doesn't state the info.

I wish it also included latency of speech to text and text to speech APIs :)

TLDR you’ll probably serve it on gpus

If it’s a small model you might be able to host it on a regular server with CPU inference (see llama.cpp)

Or a big model on cpu but really slowly

But realistically you’ll probably want to use gpu inference

Either running on gpus all the time (no cold start times) or on serverless gpus (but then the downside is the instances need to start up when needed, which might take 10 seconds)

> Apple is said to be developing its own AI server processor using TSMC's 3nm process, targeting mass production by the second half of 2025.

> According to a post by the Weibo user known as "Phone Chip Expert," Apple has ambitious plans to design its own artificial intelligence server processor. The user, who claims to have 25 years of experience in the integrated circuit industry, including work on Intel's Pentium processors, suggests this processor will be manufactured using TSMC's 3nm node.

> The Weibo user has a number of accurate previous claims, including that the iPhone 7 would be water-resistant and that the standard iPhone 14 models would continue using the A15 Bionic chip, with the more advanced A16 chip being exclusive to the iPhone 14 Pro models. These predictions were later corroborated by multiple credible sources and proved correct upon the products' release.

I compiled some - see here https://news.ycombinator.com/item?id=36195527