estebarb

I have refused to lean too hard on agentic tooling for developing. I'm aware of the gains, I use it at my daily job. But I cannot afford to loss my brain skills, just in case they do a rug pull.

These week announcements are effectively Google doing a rug pull to its customers. Now simple changes cannot be done anymore within antigravity without it to consume its full quota.

Personally I downgraded my Google One subscription. I cannot justify paying Pro anymore, and thankfully I'm not AI dependent enough to pay Ultra.

That also sucks if you are not anywhere close to retire or having a beffy bank account and depend on regular monthly payments.

No, increasing the offer of something decreases its value, always. Do not necessarily increases its demand. That is basic economic rule. See that I use "value", not "cost". The distinction matters.

Yesterday I went to a bookstore: saw an interesting book cover then I thought "ah, looks like AI"... all excitement went away. There won't be a "new complexity frontier" for artists that used to draw book covers. Or writers, actors, writers, etc.

AI is currently not enabling any use case which previously was "too hard". It is just reducing the value of stuff by increasing the offer and making people delulu about what they can achieve without proper knowledge.

Making good stuff requires paying attention to a lot of details. Even "simple" stuff can become incredible complex once you actually learn about how it must be done. Most of what we humans do is working on that space, not chasing projects Manhattans.

What do we get if population is disconnected of the true complexity of creating stuff? Perceived value decreases and if everything is perceived equally bad people will stop caring about quality. That is why fascism likes uneducated people.

So, that is about the AI contribution to "value" itself.

Now, is it true that AI will allow us to create more complex stuff that is not practical now? I would strongly disagree. The reason is Kolmogorov complexity: it is not possible to find the shortest program that describes a task. Describing it with natural language will not magically give us permission to avoid having to describe that complexity. What is the point of switching from C to English, if I still have to specify every little detail in a much ambiguous and verbose language? Programming languages are not the challenge, they are the solution to the problem of having to specify complex tasks in a reproducible way.

Now gathering everything together: that is why I think that generative AI makes things worthless: value reduction, complexity perception reduction (which reduces value), a population ignorant of the complexity will choose subpar options because "they are all the same garbage" and we will not get any superior engineering capability anyway.

It has always been possible to do it. LLMs are not a particular enabler for that.

The difference is that now it is worthless: there is no learning, no person caring about the result, nothing aspirational for the public to look towards... we used to enjoy those challenges, used to be proud of solving complex problems... now? Yeah, whatever, execute execute commit push, let another LLM "review" and call it a day.

Many banks already require monthly or annual payments for keeping an account with them. They also use the money from deposits to lend it at high interest rates. It is not like the banks are not extracting much more than a fair share of revenue from a captive market.

It is not only "their" population. Mastercard and Visa captures a % of each sale done globally with their cards. It is perfectly reasonable for all countries to want to develop their own payment systems and stop paying taxes to the USA.

Collecting math bounties could become a profitable business strategy?

Personally I would suggest that the "easiest S3" would be simply using NFS. You can get replication with RAID.

S3 is simple for the users, not the operators. For replicating something like S3 you need to manage a lot of parts and take a lot of decisions. The design space is huge:

Replication: RAID, distributed copies, distributed erasure codes...

Coordination: centralized, centralized with backup, decentralized, logic in client...

How to handle huge files: nope, client concats them, a coordinator node concats them...

How will be the network: local networking, wan, a mix. Slow or fast?

Nature of storage: 24/7 or sporadically connected.

How to handle network partitions, pick CAP sides...

Just for instance: network topology. In your own DC you may say each connection has the same cost. In AWS you may want connections to stay in the same AZ, use certain IPs for certain source-destination to leverage cheaper prices and so on...

At this point the protests should be against the matches themselves. But let's be honest: nobody cares anymore.

I didn't understood "your five-hour usage" I thought plans were per interaction or per token, not per hour.

I tried to put my postal code and it wrongly assumed that I live in the United States.

The zip first suggestion seems that would be really inconvenient for around 95.8% of the world population.

They reached Malta CSIRT. Costa Rica and Malta are totally different countries.

If this was in Costa Rica the appropiate way was to contact PRODHAB about the leak of personal information and Costa Rica CSIRT ( csirt@micitt.go.cr ).

Here all databases with personal information must be registered there and data must be secure.

I'm really surprised OpenAI went with LaTeX. ChatGPT still has issues maintaining LaTeX syntax. It still happily switches to markdown notation for quotes or emph. Gemini has a similar problem as well. I guess that there aren't enough good LaTeX documents in the training set.

Please read about the history of the region. This seems to be a good unbiased source, which is hard tobfind these days: https://www.theguardian.com/world/2023/nov/09/why-israel-pal...

In particular, put attention to this:

""" What happened to the Palestinians who were living there?

About 700,000 Palestinians were expelled or fled – about 85% of the Arab population of the territory captured by Israel – and were never allowed to return. Palestinians called the exodus and eradication of much of their society inside Israel the Nakba, or “catastrophe”, and it remains the traumatic event at the heart of their modern history.

Arabs who remained in Israel as citizens were subject to official discrimination. They were placed under military rule for nearly two decades, which deprived them of many basic civil rights. Much of their land was expropriated and Arab Israeli communities were deliberately kept poor and underfunded. """

Not only search: titles and description of Youtube videos are being translated. Colab UI is now in Spanish (using technical terms that make no sense).

Some people may want translation, mostly people that only speak a single language. But for most bilingual people, being forced a translation (a lower quality one), is a worse experience. I'm surprised that no one at Google has pushed back this anti-user behavior. It is like no one at Google knows more than one language.

The worst part is when traveling. Google ignores the browser settings, so it throws me Japanese or German website, even if my browser settings clearly says English then Spanish.

I'm confused. It was sarcasm?

Concentration camps where people is incarcerated without trial?

Companies producing their own propietary screws to block repairments is not a new problem.

I don't understand the European Union: they can force USB C on everyone but they cannot enforce standard screws?

I have found that prompting "translate my text to English, do not change anything else" works fine.

However, now I prefer to write directly in English and consider whatever grammar/ortographic error I have as part of my writing style. I hate having to rewrite the LLM output to add myself again into the text.

Currently it is impossible to use Netflix on screens connected via DisplayLink. Which is the only way to connect 2 monitors in older M processor MacBooks.

However, they at most would be the heuristic function of a search mechanism. A good heuristic, but heuristic at most. For search we need to identify when to abandon a path and which other entry point is promising. I'm not sure our current techniques are good for this kind of problems.

I cannot talk for OP, but I have been researching ways to make ML models learn faster, which obviously is a path that will be full of funny failures. I'm not able to use ChatGPT or Gemini to edit my code, because they will just replace my formulas with SimCLR and call it done.

It is funny that the article cannot be read without accepting marketing cookies

I'm a bit confused about the geometry. I'm not sure if the result ends up being like an fuzzy hypersphere or more like an "spiky hyperstar".

The criticisms are not strawmans, are actually well grounded on math. For instance, promoting energy based models.

In a probability distribution model, the model is always forced to output a probability for a set of tokens, even if all the states are non sense. In an energy based model, the model can infer that a states makes no sense at all and can backtrack by itself.

Notice that diffusion models, DINO and other successful models are energy based models, or end up being good proxies of the data density (density is a proxy of entropy ~ information).

Finally, all probability models can be thought as energy based, but not all EBM output probabilities distributions.

So, his argument is not against transformers or the architectures themselves, but more about the learned geometry.

I have been somewhat able to remove them with:

Do not offer me calls to action, I hate them.

No, the edited wording still conveys the same meaning. My edit was to fix another grammar typo.

The social networks are distributing such content AND benefiting from selling ads on them. Adding ads on top is a derivative work.

Personally I'm on the side of penalizing the side that provides the input, not the output:

- OpenAI training on copyrighted works. - Users requesting custom works based on copyrighted IP

That is my opinion on how it should be layered, that's it. I'm happy to discuss why it should be that way or why not. As I put in other comment, my concern is that mandating copyright filtering o each generative tool would end up propagating to every single digital tool, which as society we don't really want.

Not really. Youtube is not liable as long as they remove the content after a copyright complain and other mechanisms.

The problem is if OpenAI is liable for reproducing copyrighted content, so will be other products such as word processors, video editors and so on. So, as society where we will put the line?

Are we going to tolerate some copyright infringement in these tools or are we going to pursue copyright infringements even in other tools as we already got the tools to detect it?

We cannot have double standards, law should be applied equally to everyone.

I do think that overall making OpenAI liable for output is a bad precedent, because of repercusions beyond AI tools. I'm all fine with making them liable for having trained on copyrighted content and so on...