HardCodedBias

If you believe another thread the benchmarks are comparing Gemini-3 (probably thinking) to GPT-5.1 without thinking.

The person also claims that with thinking on the gap narrows considerably.

We'll probably have 3rd party benchmarks in a couple of days.

What? The 4.5 and 5.1 columns aren't thinking in Google's report?

That's a scandal, IMO.

Given that Gemini-3 seems to do "fine" against the thinking versions why didn't they post those results? I get that PMs like to make a splash but that's shockingly dishonest.

Big if true.

I'll wait for the official blog with benchmark results.

I suspect that our ability to benchmark models is waning. Much more investment required in this area, but what is the play out?

Bard was horrible compared to the competition of the time.

Gemini 1.0 was strictly worse than GPT-3.5 and was unusable due to "safety" features.

Google followed that up with 1.5 which was still worse than GPT-3.5 and unbelievably far behind GPT-4. At this same time Google had their "black nazi" scandals.

With Gemini 2.0 finally had a model that was at least useful for OCR and with their fash series a model that, while not up to par in capabilities, was sufficiently inexpensive that it found uses.

Only with Gemini-2.5 did Google catch up with SoTA. It was within "spitting distance" of the leading models.

Google did indeed drop the ball, very, very badly.

I suspect that Sergey coming back helped immensely, somehow. I suspect that he was able to tame some of the more dysfunctional elements of Google, at least for a time.

It seems like there is a very strong correlation between identity politics and "AI skepticism."

I have no idea why.

I don't think that the correlation is 1, but it seems weirdly high.

"12k a year out of your paycheck for an advanced Clippy "assistant"? Sorry, this is not plausible"

It's certainly possible. Will it actually happen? IDK.

I don't think it requires robots. Although that's possible too.

I think that HoloLens has a reasonable demonstration of how to assist blue collar work about 10 years ago (AFAIK it flopped). I would bet a dollar that similar technology augmented with LLMs could be useful to blue collar work.

This is clearly aggressive.

Will this pan out? We don't know, no one knows. But this isn't "a scam" there is a plausible future where a large percentage of white collar (or dare I say it, blue collar) work will have an assistant and that assistant requires a considerable subscription (200/mo? 1000/mo?).

Interesting to see all of the leading labs in the West make this bet.

LeCun has been talking against the company's direction, in public, for a couple of years now.

He's a great researcher, but that's abysmal leadership. He had to go.

If he gets funding (and he probably will) that's a win for everyone.

I suspect that if GenAI starts to make content which can grab people's attention, and do it cheaply, then Netflix will become far more accommodating very quickly.

They do not want to be disrupted.

Update:

It looks like there has been some turnover at the universities since the quote from Eric Weinstein. I quick survey (could be incorrect!) says it is now more around 63.

That's still high but not as high as what Eric Weinstein was quoting.

Would be a great time series.

"One example of such an "iron grip" from my colleague Eric Weinstein: Of the 67 top research universities in the US, 62 have Baby Boomer presidents (three are Silent Generation and only two are Generation X). Today, the median age of these 67 university presidents is 65 years-old... And this is very different from the recent past. Only thirty years ago, in 1990, the median age of these same university presidents was a much lower 52-years old;"

Got to admit that's very interesting. That was January 5, 2020, I wonder how it looks today.

I asked chat:

Boomers: ~13 of 20 (~65%) Examples: Princeton (Christopher Eisgruber, 1961), MIT (Sally Kornbluth, 1960), Harvard (Alan Garber, 1955), Duke (Vincent Price, 1957), Brown (Christina Paxson, 1960), Johns Hopkins (Ron Daniels, 1959), Columbia’s acting leaders across 2024–25 were also Boomers.

Gen X: ~7 of 20 (~35%) Examples: Stanford (Jonathan Levin, 1972), Yale (Maurie McInnis, 1966), Dartmouth (Sian Beilock, 1976), Rice (Reginald DesRoches, 1967), Vanderbilt (Daniel Diermeier, 1965), WashU (Andrew D. Martin, 1972), Notre Dame (Rev. Robert A. Dowd, 1965).

Millennials/Silent: 0 in this Top-20 (today). (A few large publics just outside the USNWR top-20 have Boomer or Gen-X chancellors as well; e.g., UC Berkeley’s Rich Lyons, 1961.)

The boomers are still holding on to power. Amazing!

What's it like in the private sector?

Gen X: ~53% (Microsoft/S. Nadella 1967; Alphabet/S. Pichai 1972; Amazon/A. Jassy 1968; Tesla/E. Musk 1971; Eli Lilly/D. Ricks 1967; Walmart/D. McMillon 1966; Tencent/Ma Huateng 1971; Visa/R. McInerney 1975; Mastercard/M. Miebach 1968; ExxonMobil/D. Woods 1965; plus Netflix co-CEO Greg Peters 1970.)

Boomers: ~43% (Nvidia/J. Huang 1963; Apple/T. Cook 1960; Saudi Aramco/A. Nasser 1958; Broadcom/Hock Tan 1951/52; TSMC/C.C. Wei 1953; JPMorgan/J. Dimon 1956; Oracle/S. Catz 1961; plus Netflix co-CEO Ted Sarandos 1964.)

Millennials: ~5% (Meta/Mark Zuckerberg 1984.)

Silent Generation: ~5% (Berkshire Hathaway/Warren Buffett 1930 — slated to hand CEO role to Greg Abel, b. 1962, at year-end 2025.)

A little better, but my gosh they are really holding on. I'm sure it is unprecedented. Certainly heath is better than any time in history, but this seems extreme.

"I actually think there should be a human right to data."

Interesting tagline, but probably has far too many side effects.

If it were me, I would try to boil this down to some negative right, those usually have less side effects, and even then I would be very careful.

I get almost all of my news from LLMs.

I scan the top stories of the day at various news websites. I then go to an LLM (either Gemini or ChatGPT) and ask it to figure out the core issues, the LLM thinks for a while searches a ton of topics and outputs a fantastic analysis of what is happening and what are the base issues. I can follow up and repeat the process.

The analysis is almost entirely fact based and very well reasoned.

It's fantastic and if I was the BBC I would indeed know that the world is changing under their feet and I would strike back in any dishonest way that I could.

IIUC there are few "prime" locations on the moon. NASA publicly named 13 specific candidate regions.

The nations will will likely use "safety zones" to exclude others from their base of operations. We'll see the radius of these zones but expect 200m - 2km for a start.

There is a reason to think that there is a race. Without very advanced automation all of this is pointless, but I am willing to wager that many think that advanced automation will occur within a short timeframe.

The Limits to Growth predictions are laughable.

I think that they Simon–Ehrlich wager showed how laughable they were but I guess we have to revisit every couple of decades.

Automation increases productivity.

Without automation we would all be living in poverty.

Amazon culture is entirely powered by recruiting.

Even L6 managers feel this, but it becomes more clear as one goes up in levels. Recruiting is job one.

If Amazon runs out of recruitable engineers (unlikely, they are one of the most prestigious firms in the world) then they will simply lower the bar. HC must be filled.

The company is well structured, it will survive.

Many people have difficulty processing or even finding the information on the polices of candidates. It seems reasonable to use LLMs to get that information and summarize so that the individual voter can process it.

I have no problem with people deciding, on their own, how much help they want/need to make their voting decision.

Newspapers in the Netherlands give endorsements.

My doctor requested one for me, and I just got it in the mail (USA), which is nice for me.

I suspect that this will remain a niche product. It would, in theory, be applicable to most people on earth. Such wide spread screenings simply aren't viable in a world of massive compliance costs and subsidized health care.

It took an act of god for simple COVID testing to become somewhat inexpensive.

I look forward to getting this test every few years until it is killed by our regulatory apparatus.

I agree, it is always good to be openly hostile to the new boss!

I don't know what you think "direct access to specialists" is.

I have concierge medicine. I have two specialist appointments scheduled both take about 3mo.

I can see my PCP within 1 day. That is good. I can have blood drawn within 1 day. That's good.

Specialists, no advantage. This makes it not overly valuable, but what do you expect for 8k extra for year (on top of very good health care)?

I don't know how to access a higher tier of health. Perhaps at 100M+ of net worth it appears. IDK.

There are many routes that the government has to court order/warrant/subpoena information from tech companies.

The tech companies have just about zero ability to resist.

There should likely be legislation enacted that raises chat logs to the level of psychotherapist-patient privilege.

Being a "multi-millionaire with access to the best healthcare" in the US means that you sit in the same queues as everyone else.

The best you can do is concierge care, but that only expedites primary care everything in the US is about specialists.

Of course the relation is minimal if it exists at all.

Stock buybacks are simply a more tax efficient dividend.

Maybe they don't fix it, but I suspect that they move us towards it occurring less often.

I'm sure that there were individual press releases that outlined each of these investments but it is nice to have them all in one place.

I thought that Google was doing more than ~30B in data centers, and some of the write ups talk about over the next 2 years.

I don't understand the objective if it isn't comprehensive.

"I suspect that people who are good engineering managers will also be good at 'vibe coding'."

I have observed that those who have both technical and management experience seem to be more adept (or perhaps willing?) to use LLMs in the daily life to good effect.

Of course what really helps, like in all things, is conscientiousness and an obsession for working through problems (if people don't like obsession then tenacity and diligence).

That title is silly.

The US is #2 in the world in EVs behind China. Tesla leads the US in EVs, and is #2 in the world.

A sample of more reasonable titles:

1. US is #2 in EV manufacturing and sales, can they catch up to China?

2. How did BYD catch up so quickly to Tesla?

3. Why can't Europe create its own Tesla?

Clearly:

https://pbs.twimg.com/media/DmxCKbfV4AAA59g?format=jpg&name=...