The GF3 Ti200 isn't obsolete rubbish, heck its DX8 compliant, and DX8 is the latest iteration of DX! Its as good as saying an Athlon XP 1500+ is obsolete rubbish...comon, not everyone has megabucks to spend upgrading on hardware components every 3 hours. The GF3 Ti200 will do just fine with most, if not all of today's games. Hell I'm still on a Voodoo3, which at this point is obsolete, but still ain't rubbish

A quick primer on GeForce versions
GeForce256 SDR - the original GeForce(codenamed NV10) that nvidia proclaimed will "change the world". The first consumer card to have T&L support. Core clockspeed 120MHz, memory speed 166MHz. Fillrate of 480GPixels/s
GeForce256 DDR - NV10 with DDR-SDRAM
GeForce2 GTS - 2nd GeForce iteration(codenamed NV15), GTS == GigaTexel Shader. Featured nvidia's NSR(nvidia Shading Rasterizer) and per pixel shader. Core clockspeed 200MHz, memory speed 166MHz DDR. Fillrate of 800MPixels/s or 1.6GTexels/s. Memory bandwidth of 5.3GB/s
GeForce2 Ultra - GeForce2 clocked at 250MHz, memory at 230MHz DDR. Fillrate of 1GPixels/s or 2GTexels/s. Memory bandwidth of 7.36GB/s.
GeForce2 Pro - GeForce2 clocked at 200MHz, memory at 200MHz DDR. Fillrate same as GTS, memory bandwidth at 6.4GB/s.
GeForce2 Ti - Rehashed GeForce2 with a new Ti(titanium) designation. Basically a GeForce2 clocked at 250MHz, memory at 200MHz DDR. Fillrate same as Ultra, memory bandwidth same as Pro.
GeForce2 Ti200 - Exactly the same as the GF2 Ti, AFAIK. I don't know why they came up with both the Ti and the Ti200.
GeForce3 - First GPU to include pixel and vertex shaders(codenamed NV20). Featured nvidia's nfiniteFX? Engine and Lightspeed Memory Architecture(LMA) for full DX8 compatibility. Core clockspeed 200MHz, memory speed 230MHz DDR. Fillrate 3.2Bsamples/s, memory bandwidth 7.36GB/s
GeForce3 Ti200 - First GeForce3 with Ti designation.Core clockspeed 175MHz, memory speed 200MHz DDR. Fillrate 2.8Bsamples/s, memory bandwidth 6.4GB/s
GeForce3 Ti500 - GeForce3 Ti clocked at 240MHz, memory speed 250MHz DDR. Fillrate 3.84Bsamples/s, memory bandwidth 8GB/s
GeForce4 Ti4200 - Latest iteration of the GeForce family(codenamed NV25). Features nfiniteFX? II Engine, Accuview Antialiasing? techniques and LMA II. Includes dual vertex shaders. Core clockspeed 250MHz, memory speed 250MHz. Fillrate 4Bsamples/s, memory bandwidth 8GB/s
GeForce4 Ti4400 - GeForce4 clocked at 275MHz, memory speed 275MHz. Fillrate 4.4Bsamples/s, memory bandwidth 8.8GB/s.
GeForce4 Ti4600 - GeForce4 clocked at 300MHz, memory speed 325MHz. Fillrate 4.8Bsamples/s, memory bandwidth 10.4GB/s.
GeForce2 MX SDR/DDR- First GF2MX iteration(codenamed NV11). Essentially a GF2 with half the pixel pipelines and memory bandwidth. DDR memory implementation was only 64bit. Core clockspeed 175MHz, memory speed 166MHz SDR/DDR. Fillrate of 350Mpixels/s or 700Mtexels/s. Memory bandwidth of 2.7GB/s.
GeForce2 MX200 - An even more crippled GF2 MX with the same core/memory clockspeed, but using 64bit SDR memory, effectively halving its already anemic memory bandwidth to 1.3GB/s
GeForce2 MX400 - A slightly higher clocked GF2 MX. Core clockspeed 200MHz, memory speed same as GF2 MX. Fillrate of 400Mpixels/s or 800Mtexels/s, memory bandwidth same as GF2 MX.
GeForce4 MX420 - New GeForce MX iteration(codenamed NV17). Same number of pipelines, but higher clockspeed(250MHz core, 166MHz SDR memory). Featured nview, LMA II and Accuview AA, but lacked pixel/vertex shaders of the nfiniteFX? Engine(still uses NSR of the GF2 era)
GeForce4 MX440 - GeForce4 MX clocked at 275MHz core, memory at 200MHz DDR.
GeForce4 MX460 - GeForce4 MX clocked at 300MHz core, memory at 275MHz DDR.
Then there's the GeForce Go mobile GPUs and Quadro workstation GPUs that I won't go into...