How to install tensorflow-gpu from anaconda?

I have run some very basic steps (tensorflow-gpu is currently at 2.12.1):

conda create --name py311_tf212 python=3.11 numpy numba scipy spyder pandasconda activate py311_tf212time conda install -c conda-forge tensorflow-gpu

After 3 hours of thinking and printing a few thousand lines of package dependencies, the installation fails.
My system features are:

• Ubuntu 18.04,
• feature:/linux-64::__glibc==2.27,
• CUDA 11.8,
• Nvidia driver 520.61.05.

I'm not sure what other information is relevant. I'd be happy to get any tips.

Edit 2023-12-19, I gave this a new go. This time I specified the conda-forge channel at the env creation step:

$ time conda create --name py311_tf2_test -c conda-forge python=3.11 numpy numba scipy spyder pandas tensorflow-gpu...Found conflicts! Looking for incompatible packages.This can take several minutes.  Press CTRL-C to abort.failed...Package expat conflicts for:numpy -> pypy3.9[version='>=7.3.13'] -> expat[version='>=2.2.9,<3.0.0a0|>=2.3.0,<3.0a0|>=2.4.1,<3.0a0|>=2.4.8,<3.0a0|>=2.4.9,<3.0a0|>=2.5.0,<3.0a0|>=2.4.7,<3.0a0']scipy -> pypy3.9[version='>=7.3.13'] -> expat[version='>=2.2.9,<3.0.0a0|>=2.3.0,<3.0a0|>=2.4.1,<3.0a0|>=2.4.8,<3.0a0|>=2.4.9,<3.0a0|>=2.5.0,<3.0a0']pandas -> pypy3.9[version='>=7.3.13'] -> expat[version='>=2.2.9,<3.0.0a0|>=2.3.0,<3.0a0|>=2.4.1,<3.0a0|>=2.4.8,<3.0a0|>=2.4.9,<3.0a0|>=2.5.0,<3.0a0']spyder -> python[version='>=3.12,<3.13.0a0'] -> expat[version='>=2.5.0,<3.0a0']...Package typing_extensions conflicts for:numba -> importlib-metadata -> typing_extensions[version='>=3.6.4']spyder -> ipython[version='>=8.12.2,<9.0.0,!=8.17.1'] -> typing_extensions[version='>=3.10|>=3.10.0|>=3.7|>=3.6.4']...The following specifications were found to be incompatible with your system:  - feature:/linux-64::__glibc==2.27=0  - feature:/linux-64::__unix==0=0  - feature:|@/linux-64::__glibc==2.27=0  - feature:|@/linux-64::__unix==0=0  - numba -> libgcc-ng[version='>=10.3.0'] -> __glibc[version='>=2.17']  - numpy -> libgcc-ng[version='>=10.3.0'] -> __glibc[version='>=2.17']  - pandas -> libgcc-ng[version='>=10.3.0'] -> __glibc[version='>=2.17']  - python=3.11 -> libgcc-ng[version='>=11.2.0'] -> __glibc[version='>=2.17']  - scipy -> libgfortran-ng -> __glibc[version='>=2.17']  - spyder -> ipython[version='>=8.12.2,<9.0.0,!=8.17.1'] -> __linux  - tensorflow-gpu -> tensorflow==2.15.0=cuda118py39h5387621_0 -> __cuda  - tensorflow-gpu -> tensorflow==2.6.2=cuda111py37hf54207c_2 -> __glibc[version='>=2.17']Your installed version is: 2.27real    32m40.392s

I'm not sure how to interpret this, is the installed glibc version an issue here? Most packages seem to be happy with 2.17 or newer. Oddly tensorflow-gpu has dependencies tensorflow==2.15.0 and tensorflow==2.6.2.