HardCodedBias

A person from Google famously put on her linkedin that her job was to optimize SVG for Gemini 3.0.

It's great that a novel math proof was created.

But this is mostly marketing, pleasing the sneering class/the elites who believe that simply providing value for others (through sales) is repugnant and beneath them.

It seems that these tools can do real work, and people are paying for that. IMO, that is more than sufficient.

Google Brain invented the transformer.

GDM ... not so much.

This is the correct take.

It's an unbelievable failure.

Big if true. If so they have mogged Google, and GDM in particular very, very badly.

Google Deepmind has failed.

Flash 3.5 seems capable for a flash model, Antigravity seems like a reasonable harness. But GDM is responsible for the frontier model and it looks like a complete failure.

What's particularly galling is the size of funding of GDM. It is enormous compared to the other labs. The headcount of other labs is swollen by infra, marketing, sales, GDM is pure "engineering" and its frontier model isn't even leading open source.

What a failure. It's unreal.

"I was hopeful that they avoided the well known sultry voice this go around, but alas"

Why do you care? You can select other voices. Why do you need to control others?

What is at the root of your need for domination?

While this is likely very useful to an enormous number of people, I suspect it will be even more useful for the elderly (if somehow it can be made accessible to them).

IIUC the literature, there is serious loss of functionality associated with lack of verbal interaction. People can say "they should just talk to more people" or "more people should make time for them" but the fact of the matter is that it doesn't happen, and if this helps terrific.

The premise is flawed.

"The first chart below shows that so far there are no signs of profit margins rising outside the tech sector. This is ultimately what we are waiting for, because the value of AI companies today rests entirely on the promise that margins in the S&P 493 will eventually climb."

This is absolutely not necessary. The bull case is that AI will bring great efficiencies. The surplus profits from those efficiencies could easily be competed away by firms who have adopted AI. Those firms who do not adopt AI will have their margis crushed.

A lot of claims in the article.

IDK if the WS-15 failed or is "two decades" behind. I think we just don't know, but we do know that it is that they have achieved the ability to deliver, in mass, third-generation single-crystal nickel-based superalloys. That's a strong proof point.

As for commercial, China can/has grant a sizable portion of the C919 to domestic engine producers ( I think AECC has this contract ) that allows for a lot of capital and practice.

I would not be shocked if China demonstrates highly competitive engines in the late 2020, maybe with a few setbacks and iterations. I would also not be shocked if they started demonstrating engines with some characteristics slightly better than the Western manufactures in that time period (or maybe a little later).

Huge blow to Google.

I doubt that the money had anything to do with it.

I also doubt that the state of the technology at OAI vs. Google had much to do with it, Google is behind no doubt, but the gap is not as far as we know, insurmountable.

I suspect that this is a leadership clash. Noam was working in GDM. GDM somehow went away from coding and RSI into "world models" and that has played out very poorly. Who made that call? Who was still playing politics?

Given this is Noam the list of people that could be pissing him off is very small: Demis, Sergey (?!), a couple of VPs in GDM.

What the hell happened?

The author was so close.

"This gives us the first corollary of end-state thinking — cold but honest: every step that sits between intent and implementation will, by default, disappear."

This is likely correct.

But the author withdraws into thousands of words afterwards about the importance of gatekeeping using their current skills.

I think that is wrong. YMMV.

The LLM+Harness mostly helps with execution.

These are new products (generally) and that's a different class of problem.

It is possible that since LLM+harness helps with execution then we should see more experiments.

There are massive machines filled with reactants under high pressure and cryogenic temperatures.

It is amazing that this doesn't happen more often.

Imagine being Google and paying billions to GDM just to get mogged.

Oh boy.

GDM is making (or has been backed into a corner into making) the bet that high throughput, low latency, low capability models are the path forward.

That probably works for vibe coded apps by non-practitioners.

I suspect that practitioners/professionals will wait longer for better results.

The following dates for the start of the statue of limitations were the most plausible:

1. 2019 capped-profit restructuring + 1B MSFT investment

2. 2023 Microsoft expansion / reported 75%-then-49% economics

3. 2024/2025 PBC restructuring

AFAIK it has not been reported as to exactly what the jury found, but IIUC the 2019 date is consistent with their findings.

That's poor for Musk, but it makes sense. He was arguing 2023. I think it is a valid argument.

But he had to know that 2019 was very much in play (and is likely the most logically consistent).

This is very squishy law.

I think that your first statement is correct.

I do not think that your second was correct.

I agree with the fact that your company has quota and allocates some to you.

Gemini will conduct seconds to minutes of work before requesting aid. And it will commonly fall over.

Claude/Codex will commonly do minutes to hours of work.

The difference is one to two orders of magnitude. It is immense.

He was entirely correct.

He made a follow up after the pushback by GDM.

Google’s businesses are very broad and durable. But Google being the only company in the world without access (except for GDM+labs) to a competent coding agent will take a toll.

We’ll see how long Google can hold out hoping for GDM to create something that is competitive.

I’m guess that within 6 months Google will give up on coding and finally let their devs use Claude/Codex.

This isn’t a security problem, this is a GDM issue with GDM’s promises being far beyond their ability.

This one is different? What about unitree? What about their demo at the Spring Festival Gala?

https://www.youtube.com/watch?v=Ykiuz1ZdGBc

That sure felt "different".

No doubt hands are important, but I think you've missed a lot here Wired.

I use ChatGPT all of the time, but the model backing the voice model (or it's settings) is intensely stupid.

If Grok is actually good here, they will have a customer!

You are entirely incorrect.

Here is the LLM's summary of the current legal issue at hand:

Attempting to determine the identity of an unknown individual co-located with a victim at a specific time requires a reverse-location query. Because the Supreme Court has not yet established a unified national doctrine for these searches post-Carpenter, lower courts are highly fragmented. Many magistrates systematically refuse to authorize geofence warrants or tower dumps, citing the lack of individualized probable cause for the peripheral, innocent devices swept up in the geographic net.

And indeed, in my case, the police were not able to conduct this geofenced investigation (which would have instantly idenitied the person).

Most people don't understand how powerless police are to find criminals. That they catch them at all is often amazing. I have firsthand knowledge of this from a tragic loss in my family. The investigation was severely hindered because investigators could not utilize cell location data, despite knowing someone was present at the scene. Police spent an extensive amount of time trying to identify them without success. When the identity was eventually discovered through entirely different avenues, it confirmed the individual had a cell phone on them. The location data would have resolved the identification trivially. We should enable this capability and put strict "guardrails" on its use.

Everyone at Meta should know the score.

Meta pays top dollar. They also pay enormous sums for what management identifies as performance.

Conversely, Meta is ruthless about cutting those management identifies as low performers.

This is the deal going in. It’s not a crime.

This is an amazing technological achievement. Kudos.

This is so much better than the competition. I suspect that this will have an impact in business and education, at a minimum.

Impressive Tenue, IMO.

Apple Watch, AirPods, M1 Silicon, services.

A few flops, like Apple Vision Pro and their confusion with AI. But that's ok given the wins.

Overall, as a non-founder he's near the tops in CEOs over the last couple of decades. The only non-founders I would put above him are Satya (although he has a had a couple of rough years), Bob Iger, Jamie Dimon and maybe Andy Jassy.

Taking a fair lens to this he is "first round hall of fame non-founder".

"As of two years ago they were still liquidating out of it"

I get that people are scared of investing in China. But if I still made single stock investments, I would seriously consider BABA, it seems well positioned.

I think its clear that a lot of companies over hired during the pandemic, and it seems like we are getting more out of our entire staff with AI tools.

Given that it makes sense to cut if the company doesn't have new product ideas to spend that excess resources.

In general:

Cynicism makes you sound smart. Optimism makes you successful.

The cynicism around this technology is everywhere, even though it clearly has real power to solve problems. It is a technology which enables so many use cases that were impossible before, that makes it very highly hyped/expected. And that is causing an immune (over) reaction by natural skeptics, that's an error.

People need to take a measured, reality based, view of how the technology is being used today, the adoption curve, and the increase in capabilities over time.

It's clearly being used strongly, and may even be revolutionary.

Bubbles burst when there's no 'there' there. AI has an undeniable 'there'—the only question is the timing of the ROI.

The power draw is going to be crazy (today).

Practical LLMs on mobile devices are at least a few years away.