mrbungie

It is a feelpiece, but I don't think that's necessarily a bad thing, at least it invites discussion.

Also, framing it as "us vs them" misses the point. For instance, in my world (Data & Analytics), there's no universe where Spark makes sense for <1TB data or using Kafka for what could be solved using with more simple and predictable batch processing, yet we still see them used in wild implementations that clearly don't need that scale. Those implementations can end up needing a lot of headcount and budget for support without adding any value whatsoever and sometimes even ending up in a more brittle platform. Complexity often comes from misguided choices, not just ego.

I've seen enormous data teams at startups, by that I mean bigger than WhatsApp engineering team in its prime pre-Meta (~30 people). That's a red flag, especially when they're processing far less data.

Afair (I might be wrong) AWS and a big chunk of the industry is promoting Iceberg over Delta. Delta is mostly backed up by Databricks.

What I understand from what GP was saying, is that AI Alignment today is more akin to trying to analyze and reduce error in an already fitted linear regressor rather than aligning AI behaviour and values to expected ones.

Perhaps that has to do with the fact that aligning LLM-based AI systems has become a pseudo predictable engineering problem solvable via a "target, measure and reiterate cycle" rather than the highly philosophical and moral task old AI Alignment researchers thought it would be.

Looks very cool! I will take a look, tysm!

Yeah, it just would be great if it already did so and I hope it supports Iceberg soon, as it would enable me to change expensive (and bad) engines like AWS Athena for something more manageable.

Don't get me wrong, I'm just being a tongue-in-check egotistical bastard data engineer from hell. DuckDB is a fine piece of software as it is, and those mantainers deserve heaven.

I was almost going to build a lakehouse* with DuckDB because I low-key love it, easiest and strongest analytical engine I've found yet: scale from laptops to big metal, while being mostly out-of-core when doing sane stuff, and avoiding distributed computing for SQL in the process (looking at you Spark).

That is until I found out it does not support Iceberg writes[1], big nono as I would need another engine for inserts, and I want a simple stack :(. What a bummer.

[1] https://github.com/duckdb/duckdb_iceberg/issues/37

*that is what they are called now aren't they? I just can't follow the terms anymore haha.

He's very good at doing startup/scaleup management, in a ruthless/snakey skill level. I think that's something everyone can agree.

But vision? I'm not so sure, he had great company and help along the way, and now that he has been left alone (arguably due to his own actions) he's selling the image of competence in areas that he hasn't demonstrated skills whatsoever. We'll see.

> It seems that should be able to replace 1st level support for a bank or something like that but those industries move very slowly so even that could be years away still.

Klarna is similar to a bank and they tried replacing 1st level support with a chatbot. Spoiler: it didn't work for them, so I wouldn't keep my hopes so high for banks.

All that YC/VC experience paid off, they're the masters of gutting things from the inside out in the name of growth.

Sure, but at least it would've been tethered to someone who actually knew him. Remove that and any humanity in the quote banishes.

You know you are in front of the impending explosion of a bubble when discussions shift from products themselves and towards who will be working with whom.

If Ive had said that himself, it wouldn't be weird. But Sama didn't know Jobs well enough to put words in his mouth.

Narcissistic ego stroking. He literally is invoking and thinking out loud for a dead man, who was not that intimate with him, just to validate his own ventures and acquihires. Very weird.

It wouldn't be that weird if Ive had said so himself.

Wow, such a uncalled straw man out of nowhere.

That's common when they are distributed as single notebooks (i.e. via Google Colab). When distributed inside repos they usually contain a requirements.txt.

Run of the mill corpo speak.

It's a bit more impressive when you consider the size of these buddies.

Are you sure about that? I'm pretty sure the 16" models from 2019 also had a touch bar, although as you said the keyboard was much better.

Pieces of shit. And then they assign you a score for each travel, as if you are really "carpooling" when in reality is a shitty taxi replacement (not that taxis are on a moral high ground, but the point still stands).

I had (unstable) cellular service all the way through, and chatted with other people too. Probably commune/zone-related.

Yep. The 2010 Chilean earthquake and tsunami[1] is an example of what can happen in this country during catastrophes: sacking and looting.

Since then night curfews are a given during and after disasters.

[1] https://en.m.wikipedia.org/wiki/2010_Chile_earthquake

Most telco and central gov. tech infra was actually working, unstable, but working during the first hours of the outage. I know some friends eventually lost connectivity in remote parts of the country though (for 2-3 hours at most).

For the average chilean: As long as you had 4G-5G equipment with working batteries, you had internet and access to essential gov services.

Chilean here. The curfew was very reasonable due to the scale of the outage initially affecting the ~99.0% of homes, and its the purpose was just to reduce circulation due to restricted policing capacity and the inherent danger of big cities such as Santiago being completely out of energy and artificial light at night.

Plus, both natural persons and enterprises could ask for curfew passes via a website (Comisaria Virtual) created during the COVID-era for doing police paperwork online. Military/police staff were especially understanding and people who didn't have a pass when checked were allowed to get one on-site using their smartphones.

It was as soft as a curfew can get. I didn't see any important backlash on social networks if that says anything to you.

I really doubt the technology would've advanced as fast without such an adversarial setting. Economic warfare is child's play vs total war.

All that knowledge and redirected output ended up being extremely profitable once it could be used for civilians instead of the military when the war ended.

TSLA is a meme stock. PE ratio vs Sales Growth is ridiculous vs the other Mag 7.

Bill Gates dedicated to white-washing his image after MSFT antitrust case.

I mean, philanthropy is great, even more if it is sincere. But lets not forget the kind of business man he was before the 2000s.

Unless dealing with a psychopath you can deal with the lies using other subsystems.

Ask Musk how Tesla sales in EU are doing once enough people started distrusting him directly as a person.

Economic votes are based partly on trust, and with enough swinging in the distrust direction, it would have a direct repercution on OAI financials.

It has served us relatively fine for thousands of years.

LLMs? I'm waiting for one that knows how not to say something that is clearly wrong with extreme confidence, reasoning or not.

Give a group of "average human" two years, give or take 6 months, and they will also saturate the benchmark and probably some humans would beat the SOTA LLM/RLM.

People tend to do so all the time, with games for example.