"Not thinking much of people" is the right take not because it is mean, but because of reality.
Humanity is something like a great inert blob. It takes tremendous effort and charisma and trickery and cajoling and organizing and "nudges" to get the blob to move. Fear is a great motivator, but if we do not fear, we cannot be made to move.
There is no such thing as a spontaneous uprising in humanity. The flash mob or corner protest was organized. Fads like the 6ft social distancing floor stickers were printed and placed everywhere practically overnight and done with terrific social engineering forethought.
I tell people 2027 is going to be the year of deferred tech debt. But sometime afterwards, all that tech debt will have been addressed, and new software will be run through the same models hacking the old software. The worst is yet to come, sure, but then it gets better when the models run out of broken systems to attack.
it seems a race between models getting better and tech debt accruing as well as companies wanting to spend the tokens on paying that down vs new features: I was very much into the camp of "in a few years all these chickens will come home to roost" but now I am not so sure.
I still think right now the cyborg approach is the best (like it used to be for a while in chess), even fable 5.2 needs steering, but the amount of steering I have to do now vs the amount of steering I had to do a year ago is significantly down. If the rate of improvement remains I can see fable 8 being able to be told "make this codebase more maintainable so we can fix the tech debt" with not a lot of supervision, and if we get there all bets are off on the future of coding as a career.
So open weight models are going to eat the world even if we pass a domestic ban on them in the USA. And that means the world is at worst 3-6 months behind Anthropic and OpenAI firing on all cylinders burning mountains of cash to stay ahead. Their inference business can likely support that, but there is no moat, just a 6 month lead that could evaporate at any time.
So then the US restricts HW sales. Cool, the world makes it up in volume by running distributed inference or training just like Deepseek did. Components are expensive right now, but that, much like the tech debt, will ultimately be addressed.
There are apparently ~50 edge computing startups and ~100 AI HW startups right now. I would have said none of them stood a chance against Nvidia, but now that the federal government has commenced punching itself in the face over AI as they attempt to stop an avalanche by grabbing pebbles, anything goes 3-5 years from now when Nvidia's current roadmap runs out. And it's a lot worse than just the loss of the CEO, it's all the wasted energy by their staff jumping through hoops to prove to a bunch of soundbite chasing technologically ignorant Luddite politicians once again that Nvidia is not the Enron of technology. At worst, they were playing by the letter of the law as the violated the spirit, just like every single tax-dodging multimillionaire does annually.
None of this leads to WW3, but I think it leads to us falling behind for good. And given our choice of leadership and our geopolitical shenanigans, that's probably for the best.
I checked out of inequality when California voted down the billionaire tax. Apparently, the past two years did not established a plurality that believe there should be a tax on unrealized gains if they are used to secure financing or deployed in any way.
But I think that recession is a phase transition for which we are deers in the headlights of the future. Unfortunately, we do not have leaders even trying to fix the tax code or manage UBI funded by some sort of tax on the runaway profits of AI if they emerge at all. So we'll fix it in the most expensive way when we run out of alternatives, as usual.
Sure, chance favored the prepared processor. But given the number of times the graphics people tried to kill cuda in advance of that, without his absolute control, I don't think it would have survived to its day in the Sun.
And you can see that at AMD and at Intel with all of their various hardware and software alternatives that came and went in the same timeframe because no one stood up for them.
Those companies treated those projects as research. Jensen treated cuda as a product and immediately started figuring out how to make money with it.
On top of that, hasn't intel now end-of-life'd its GPU division - at least for gaming. It may still live on as a "Compute Card" in datacenters but their iGPU and gaming GPU have been effectively replaced with nvidia silicon in the roadmap I've seen.
That said, Intel has its fair share of mistakes in products - they dumped HBM before the AI boom demanded it as one of the other examples I can think of.
And the average US citizen re-elected the poster child president for "rules for thee, but not for me." And IMO if he hadn't started the Iran War, imposed tariffs, and driven inflation through the roof, his approval rating would still be 47%+ like at the beginning of his second term.
That's not true at all. Political factions routinely take steps to try to discourage their opponents from voting at all. They run campaigns on opponents' media with the goal of building cynicism. Voter registration laws are inconsistently enforced to raise barriers to certain people voting. Employers (illegally, but often) discourage employees from having the time/availability to vote if they're worried about what they'll vote for. Hell, other countries interfering in elections often do so by spreading misinformation with the goal of getting people who would otherwise vote for a given agenda to just give up.
Those things wouldn't happen if "too lazy to care" was the only reason people didn't vote. A lot of people who don't vote absolutely should be counted for a given side; they don't (or can't) vote due to reasons much more externally engineered than laziness.
> they shouldn't be counted by either side as a relevant position.
Yes, they shouldn't be, but the losing side is always count them as such. Especially if they would had lost anyway.
> failure to vote is not a "silent vote";
The "Failed to vote" and "didn't even bothered to" are two different things. And the latter effectively gives the winner more weight by denying other contenders even a chance to receive a vote for them. Hence the 'silent vote'.
I think there is a huge amount of resistance to a near-term future where it's more about the ability to describe and specify a thing than it currently is to code up a thing in order to build it. And I suspect that LLMs or whatever succeeds them will continue improving from here without any of the claptrap about a near-term malevolent superintelligence that's out to kill us all.
The future IMO will be relatively efficient code with independent models reviewing for quality and security afterwards. Life continues to come at us fast.
Programming languages are a terse way to describe the logic we want, so if you understand what needs to be built well enough it's possible to write it out faster in code.
Natural language is more ambiguous so you often need more of it to describe something. Specifications if consistent often end up as a kind of programming language in themselves, but usually ambiguous on matters that affect performance.
I'm with you on the review side though. Even the best models of today give some nonsense in reviews, but if you understand the code it's usually easy to skim the review and skip over nonsense.
I've been programming for many decades and I go all the way back to FORTRAN IV to C/C++ through CUDA. That said, there are tasks I can more easily describe than implement. I no longer bother resolving config issues on my machines, I let the coding agents do that as one example and then explain their fix to me.
But you are right about the efficiency of description given you have the 10,000+ hours into writing efficient code in the chosen language. Most do not. And they hide that by leaning on Python and its insane library toolchain which itself becomes its own skill set, but one more amenable to LLMs to master.
To be brutally efficient, one would write everything in machine code, no?
One more thing: any low-level coding legend will tell you to focus on the 2% of the code that eats 98% of time. That leaves the other 98% of the codebase as fair game for increasingly trustworthy coding models as AI becomes the new compiler.
If you micro-optimize an entire codebase yourself, not only will you never get the job done, the hardware and architecture underneath will shift faster than you can specify your precise intent. CUDA itself proved this 20 years ago with 100x to 1000x accelerations over legacy single-threaded and or badly distributed CPU code, but I guess the lesson didn't sick in.
I had my order for a bed frame from Amazon diverted to some guy's house two cities over a few months back. I'm sure that wasn't the only he "received."
They'll be fine for a few years with a talking donkey running the place because they have a roadmap. And then when the situation on the ground has changed and the future is ambiguous, they will nosedive like SGI did. My money is on Huawei now. Way to go America! You've actually outdumbed the CEO's poor choices of mistresses.
Seems more to it than that. Allegations Bill and Jensen and Rick planned it. Inside job kind of deal, with Rick being rewarded with COO job and a forgiven multimillion loan from Microsoft. https://www.youtube.com/watch?v=GW07AIaJDPM
Not at all. And if there was any smoke there back then, the 2002 microscopic accounting investigation of Nvidia would have dug into it given they were trying to portray Nvidia as the next Enron at the time. But conspiracy nuts gotta conspiracy nut so whatevs.
> But conspiracy nuts gotta conspiracy nut so whatevs.
I'm unsure of your meaning. It seems fairly clear. SGI sues Nvidia. New CEO Rick comes into SGI. CEO changes SGI's course, kills their roadmap, kills the lawsuit against nvidia, transfers SGI's patents and engineering team to Jensen. CEO suddenly quits. 3 weeks later, same CEO joins Microsoft as VP of the division that then does Microsoft XBOX which then chooses Nvidia. Microsoft gives 200 million to Nvidia. CEO Gets interest multimillion free loan from Microsoft that is then forgiven.
Please tell me what I have gotten wrong or incorrect in above facts. Is all of that considered normal business practice in your country?
Look, you do you. You are free to believe that in 2002 when the SEC conflated 6 piddly insider trades by Nvidia employees into the Enron of technology, only to all but clear Nvidia of wrongdoing beyond understating income by $10M or so and punishing all the employees that performed those trades, that somehow the SEC itself was paid off not to notice this. I choose to believe otherwise and if my memory doesn't fail me, all the purchases of fallen competitors got lumped into the investigation.
So seriously, IDGAF in 2026. Nvidia all but died during that investigation given the double whammy of MSFT shafting them by changing the DirectX 9 spec to favor ATI and the brouhaha over absolutely piddly insider trades by naive employees (except for Robert Prevett who was an over the top but ultimately amusing idiot).
I asked 2 simple yes/no questions. I guess you don't want to validate that? Please tell me what I have gotten wrong or incorrect in above facts. Is all of that considered normal business practice in your country?
I despise target retirement age funds for exactly this reason and I have all my 401k funding in major index funds that I do not care about beyond that on a day to day basis.
The major target retirement date funds are just "funds of funds" that are distributed across stock and bond index funds in an age appropriate allocation. e.g. https://investor.vanguard.com/investment-products/mutual-fun... They're generally not actively managed funds chasing bubbles.
reply