bg24

Last week, I made the explicit decision to circle through paid subscriptions of various providers every few months, instead of sticking to just ChatGPT.

Glad to see Meta.ai on the list, in addition to Claude, Gemini, Mistral and Perplexity.

"Looking holistically, our end-to-end LLM pretraining pipeline has become nearly 4x more compute-efficient in the past ten months."

I did not fully understand the technical details in the training efficiency section, but love this. Cost of training is outrageously high, and hopefully it will start to follow Moore's law.

Agree. Nitpicking just one important things. A small % of US workers can compete. That's why we see best-in-class companies born in US.

But mass workers... NO. So you are correct.

We focus on solving problems (customer pain points) faster, better and cheaper by leveraging AI, in our domain of expertise.

Unfortunately I have nothing to showcase yet.

Learning and fitness are great ways to avoid the feeling of doom.

Open-source will catch up in <6 months. Note that Meta will ship llama3 anytime. So will Mistral.

I am a PM, and switched to becoming a builder. Enjoy learning, keep building. What people take time to realize is that building things is a habit. As you combine that with ongoing learning, you will enjoy the process and eventually build something to earn a living.

I think it is a good outcome for Cilium. With the extent of open-source, it would have been difficult to grow their ARR. Good outcome imho, because Cisco is a networking company at heart and they know how to grow a network/security business.

Big fan of your products (Vagrant, Terraform, Vault, Consul). Thank you for the amazing contributions to the community. Best wishes.

Having a walled garden is fine. I use iPhone and apple products from day 1. That said, charging 30% of subscription in the App Store seems ridiculously high if you put yourself in the 3rd party vendor place.

For comparison, Amazon charges 15%. AWS marketplace charges <10%.

How is a system designed to give people the rating of “exceeds, meets, or fall below" expectations”?

I can only imagine permutations involving rapid assessments by humans (manager and team). Same as performance review, but everything (primarily exits) are executed faster.

The challenge with these repairs is that they only focus on the body part. However, nobody knows the damage to the other systems in the car when hit at 45mph. Imo, it would have been better for the insurance company to total the car and pay you to buy another one.

Agree. Or allow users to create their focus areas (tabs) by automatically categorizing topics based on the content.

Yes. And if you really needed to scale beyond 1 node, enable swarm mode.

My hypothesis is that they want to get the absolute mindshare and adoption and decide the right product/solutions along the way. Note that there are trends (web 1, web 2, mobile etc.) and the leaders tend to grow significantly.

It is a mad rush time with the new generation of AI, and future products are going to be amazing. Meta cannot afford to have just 1 cash cow. Metaverse failed, but they have a real shot here.

A few things I can think of (beyond social networks).

- Enterprise push, starting with very high-end enterprises. Their license structure (free below X million users) allows it nicely. They can afford to have a small team and yet win big. Note that typically 80% of any enterprise products revenue comes from say, 10-20% of top customers.

- Build a platform. Meta learned from Google, and then Apple debacle on how dependent the world is on platform. What would the new OS look like?

- Weaken Google and Apple. Some indications on why they partnered with Microsoft - this also ties to the point above.

I think you missed this part:

ChatGPT Enterprise is also SOC 2 compliant and all conversations are encrypted in transit and at rest. Our new admin console lets you manage team members easily and offers domain verification, SSO, and usage insights, allowing for large-scale deployment into enterprise.

I think this will have a solid product-market-fit. The product (ChatGPT) was ready but not enterprise. Now it is. They will get a lot of sales leads.

Very sad. But also a miss on admin's part, even if unfortunate. They did not realize during manual restore testing that volume was not being backed up.

"Yes and also apparently no. We use #Velero to capture backups of our cluster every 6 hours. From what I had seen our backups had been running successfully. I discovered once the incident started that backups had captured everything but the Persistent Volume Claim data. While manual backup and restore tests were run once a month to ensure our backups were functioning, they were run manually. After digging into why our restores were not coming up with data, I found that our recurring backups were missing the flag to run volume backups with Restic which snapshots PVC block volume data."

Customer acquisition cost. Lifetime value.

I am not sure about whether the metrics above are benchmarks for CAC based on advertisement alone.

I was freaked out back in Jan, but now I believe it will propel white collar jobs. We just need to wait until someone starts selling Raspberry Pi of robots, things you can assemble to do physical tasks. That will unlock a significant advancement.

Forcing engagement is in Meta’s culture. Out of the blue, suddenly I have started seeing friends notifications from my contacts list. 9 out of 10 times, people on Instagram will generally accept this. But Twitter/Threads users are different. Instagram culture may not necessarily work for Theeads.

It depends on your use case, correct? If you do not have a heavy inferencing requirement, then CPU is good enough.

Nice analogy and explanation. Another aspect is building a ubiquitous platform and figure out how to monetize later as they (Meta) already have a cash cow.

Zuck is a smart leader. Metaverse was a debacle. But the new world (AI centric) is for real. He is likely focusing on both weakening the stronghold of Google and building a massive community (like Android) around llama. Product ideas (including enterprise focus) will emerge over time.

If you work for a big corporation, that seems to be the norm. Only choice out being self-dependent.

“ A key difference between NotebookLM and traditional AI chatbots is that NotebookLM lets you “ground” the language model in your notes and sources. Source-grounding effectively creates a personalized AI that’s versed in the information relevant to you. Starting today, you can ground NotebookLM in specific Google Docs that you choose, and we’ll be adding additional formats soon.”

While it sounds promising, the blog post should have explained the obvious thing: how is source-grounding different from giving a document to a model and generate insights?

Having spent 4 years working with kubernetes (though as a PM, but pretty hands-on), getting started is easy - like in less a week. The problem happens when you run into issues. That can suck up lot of time. Also if you are new to containers, it might not be a good step to venture into k8s.

In your particular case, you have identified a problem (restaurant owners finding it difficult to keep their online menu's up-to-date).

Is the problem real? How did you find out? Talking to a few restaurant owners (who are keen to discuss) is the starting point. This will also give you insights into other adjacent problems (eg. why should I have to update my menu on my internal database and web portal?).

Is the problem worth solving? How often the restaurant owners need it? Is it important enough that they will pay for it? Do you believe you need to bill for it, or give it out for free to build the base for a future paying product?

As pointed out in the comments, go-to-market is really important. In your particular case, you have the opportunity for self-service acquisition, or via affiliates, or via partnership with Uber Eats, Yelp etc. to simplify on-boarding for restaurants. Go-to-market becomes a deal breaker as soon as you have a good product-market-fit.

Everything is defined in language. As a system is able to learn from the corpus of texts, it also builds reasoning to a large extent. For example, GPT4 can guide you through solving a problem like building a flying and crawling drone for inspecting an attic. Or, watering plants. So this goes beyond language, unless I missed your point.

Google has 3 user-facing products directly in the AI chat space: Search, Bard, and APIs. What you listed the the backend intelligence (LLMs).

For consumer segment, the products (Search, Bard) will continue to evolve. It will be interesting if there is an opportunity (eg. Search Pro).

For business segment, expect the API to evolve as they expose more powerful LLMs. In the process, they will deprecate some of the older APIs, which is expected.

In my opinion, it will be a mistake to assume that Bard and Search in the current form will be still relevant a year from now. So expect these products to be re-imagined and new product/extensions to show up.

A large percentage of employees can definitely walk out if they do not need to worry about paying their bills, should things turn south. There will be less demand and more supply of workers in the foreseeable future.

My opinion is generic across all companies. I am not an Amazon employee, nor own any stocks.

I am not an expert, but Meta might be missing out on a business opportunity by being blindsided on everything social media. Over time, every company (Google, Apple, Amazon) diversified themselves. With their AI firepower, they could have explored a number of new avenues. Every business need not boot with billion users.

I recently moved to Caddy. Out-of-the-box TLS, HTTP3, Dumbed-down simple configuration, static files... it has been a breath of fresh air.

You do not need to use search engines, unless you have to. For latest information, ChatGPT (and LLMs) are already figuring it out.

It will also have an impact on how user interfaces are being built. Why bother with so much minute design, if you can manage with a Google like search bar, and a voice button.