News for hackers. Stuff that matters.
Re:Advancing chip design with OpenAI models by epistasis on Wednesday November 18, 2026 @02:19PM (#50550756)
I've heard this too, all the IC designers I know think there's some fundamental training missing that prevents the models from creating highly optimized circuits, which is what's needed for most tasks of interest.
Re:Advancing chip design with OpenAI models by ducktective on Wednesday November 18, 2026 @03:02PM (#50551372)
How is chip design more complicated than Navier-Stokes?
Re:Advancing chip design with OpenAI models by Onavo on Wednesday November 18, 2026 @03:07PM (#50551431)
A lot trickier to do self supervised RL on chip design than Navier Stokes, or maybe they aren't trying hard enough. I mean even if you check their outputs with Altium and KiCad DRCs it won't capture everything that goes into a good board. Still need human training data which is slow and expensive.

I suspect before the end of next year they will have enough data to bootstrap a production level self supervised learning environment.

I wonder if they do training verification with real pick and place circuits too.

Re:Advancing chip design with OpenAI models by fooblaster on Wednesday November 18, 2026 @03:14PM (#50551529)
This is simply a case of they haven't gotten to it yet. They definitely will be able to do RTL optimization.
Re:Advancing chip design with OpenAI models by mathisfun123 on Wednesday November 18, 2026 @03:43PM (#50551869)
do you understand how large the training sets are for LLMs? there is not enough RTL in the world (let alone public RTL) to usefully contribute to the training of a O(T) model. and contrary to popular belief, HDLs are nothing like programming languages so nothing transfers.
Re:Advancing chip design with OpenAI models by Onavo on Wednesday November 18, 2026 @03:51PM (#50551961)
If they can figure out a good classifier for good vs bad circuits then that should be enough to bootstrap a robust model. It's just extremely extremely expensive getting good EEs to do this sort of data annotation for them.
Re:Advancing chip design with OpenAI models by mathisfun123 on Wednesday November 18, 2026 @04:07PM (#50552120)
you're just saying the same thing in different words (classifiers are trained on data)
Re:Advancing chip design with OpenAI models by Onavo on Wednesday November 18, 2026 @04:50PM (#50552672)
Not necessarily, for electronics it could be done using a shit ton of heuristics. It won't be a good classifier, but for the purposes of bootstrapping ML you can do it step by step until you get a Good Enough classifier.
Re:Advancing chip design with OpenAI models by petra on Wednesday November 18, 2026 @06:40PM (#50553925)
This isn't done using a transformer, they probably doing something else .
Re:Advancing chip design with OpenAI models by fooblaster on Wednesday November 18, 2026 @08:38PM (#50555005)
did you miss the part where they partnered with synopsis? you think they might have some reference RTL?
Re:Advancing chip design with OpenAI models by mathisfun123 on Wednesday November 18, 2026 @10:09PM (#50555569)
.... Are you implying synopsis has enough training data to provide them?
Re:Advancing chip design with OpenAI models by fooblaster on Wednesday November 18, 2026 @11:07PM (#50555964)
yeah I am
Re:Advancing chip design with OpenAI models by mathisfun123 on Wednesday November 18, 2026 @11:32PM (#50556113)
you're wrong
Re:Advancing chip design with OpenAI models by WarmWash on Wednesday November 18, 2026 @03:07PM (#50551436)
It's not that it is harder so much as it is that the training set is incredibly small.

Mathematics is generally done in the open. SOTA chip design is done by a handful of people all behind closed private doors.

Re:Advancing chip design with OpenAI models by petra on Wednesday November 18, 2026 @03:38PM (#50551823)
But in general, chip designs are built out of templatized modules with verifiable goals, and those templates are assembled into systems.

So instead of transformer based learning, they can probably combine design space exploration with some sort of understanding/generalization.

The biggest question is about efficiency, simulation time, etc...

Re:Advancing chip design with OpenAI models by acomjean on Wednesday November 18, 2026 @03:47PM (#50551914)
With proprietary synopsis software.
Re:Advancing chip design with OpenAI models by mathisfun123 on Wednesday November 18, 2026 @03:37PM (#50551797)
reminder: LLMs aren't actually intelligent, they just have trillions of parameters fit on a training set.
Re:Advancing chip design with OpenAI models by azan_ on Wednesday November 18, 2026 @05:31PM (#50553254)
Give me definition of intelligence that excludes LLMs but includes below-average intelligence person.
Re:Advancing chip design with OpenAI models by mathisfun123 on Wednesday November 18, 2026 @05:49PM (#50553480)
did you read this thread? the answer is "ability to write RTL without reading/training on 20T RTL tokens".
Re:Advancing chip design with OpenAI models by nutjob2 on Wednesday November 18, 2026 @06:25PM (#50553795)
Give me any rigorous definition of human level intelligence.

So far it doesn't exist.

Re:Advancing chip design with OpenAI models by Symmetry on Wednesday November 18, 2026 @04:00PM (#50552059)
Navier-Stokes is ultimately a textual problem rather than reasoning about shapes in the physical world like laying out gates. There are, e.g., mazes small enough that LLMs can solve them but generally they're much worse at stuff like that than humans.
Re:Advancing chip design with OpenAI models by xyzzyz on Wednesday November 18, 2026 @04:36PM (#50552457)
Navier Stokes is about modeling fluid flow through space. It’s not any more textual than chip design. Just like you can reduce fluid flow to textual partial differential equations, you can reduce chip design to purely textual constraint systems.
Re:Advancing chip design with OpenAI models by skyberrys on Wednesday November 18, 2026 @04:07PM (#50552123)
I personally think part of what makes chip design harder is a good amount of the training data is proprietary, and of training data you can find in public, over half of it is poorly designed or only acceptable for FPGA deployment not ASIC.

So you have to figure out what training data is acceptable and also make deals to get access to proprietary data.

Re:Advancing chip design with OpenAI models by fooblaster on Wednesday November 18, 2026 @04:21PM (#50552274)
That is probably what partnering with synopsis is effectively solving.
Re:Advancing chip design with OpenAI models by fsbonetto on Wednesday November 18, 2026 @06:25PM (#50553797)
I've posted a very similar research back in April (https://github.com/FeSens/auto-arch-tournament), did not make as much noise as openAI of course
Re:Advancing chip design with OpenAI models by scrlk on Wednesday November 18, 2026 @03:39PM (#50551831)
Am I missing something because DEC apparently cracked this via expert systems in the 80s?

> SID (Synthesis of Integral Design) was a logic synthesis program used to generate logic gates for the VAX 9000. From high-level behavioral and register-transfer level sources, approximately 93% of the CPU scalar and vector units, over 700,000 gates, were synthesized.

> SID was an artificial intelligence rule-based system and expert system with over 1000 hand-written rules. In addition to logic gate creation, SID took the design to the wiring level, allocating loads to nets and providing parameters for place and route CAD tools. As the program ran, it generated and expanded its own rule-base to 384,000 low-level rules.

> Initially it was somewhat controversial but was accepted in order to reduce the overall VAX 9000 project budget. Some engineers refused to use it. Others compared their own gate-level designs to those created by SID, eventually accepting SID for the gate-level design job. Since SID rules were written by expert logic designers and with input from the best designers on the team, excellent results were achieved. As the project progressed and new rules were written, SID-generated results became equal to or better than manual results for both area and timing. For example, SID produced a 64-bit adder that was faster than the manually-designed one. Manually-designed areas averaged 1 bug per 200 gates, whereas SID-generated logic averaged 1 bug per 20,000 gates. After finding a bug, SID rules were corrected, resulting in 0 bugs on subsequent runs.

https://en.wikipedia.org/wiki/VAX_9000#SID_Scalar_and_Vector...

Re:Advancing chip design with OpenAI models by epistasis on Wednesday November 18, 2026 @04:13PM (#50552203)
That sort of logic synthesis program has plenty of equivalent tools in use today, though I'm not sure if any expert systems are still used. CAD is essential in all phases.

But that's not really relevant at all to asking Claude to fix your counter, or figure out the logic and layout for a certain goal, or whatever.

Re:Advancing chip design with OpenAI models by maherbeg on Wednesday November 18, 2026 @04:46PM (#50552598)
All it takes is to offer some retired IC designers a ton of cash to enjoy their retirement more and provide some solid training data and verify some outputs every week.
Re:Advancing chip design with OpenAI models by jacquesm on Wednesday November 18, 2026 @06:53PM (#50554046)
'some solid training data' -> an example or two would be nice. Otherwise 'some' might turn out to be 'quite a lot more than expected'.