From 52e3698cddcbf3c638990196cbe0d6c7aa81bb0a Mon Sep 17 00:00:00 2001 From: Yanir Seroussi Date: Mon, 8 Jul 2024 14:09:24 +1000 Subject: [PATCH] New post: Exploring an AI product idea with the latest ChatGPT, Claude, and Gemini --- .../chatbot-arena-leaderboard-2024-07-08.webp | Bin 0 -> 37692 bytes .../index.md | 135 ++++++++++++++++++ 2 files changed, 135 insertions(+) create mode 100644 content/posts/2024-07-08-exploring-an-ai-product-idea-with-the-latest-chatgpt-claude-and-gemini/chatbot-arena-leaderboard-2024-07-08.webp create mode 100644 content/posts/2024-07-08-exploring-an-ai-product-idea-with-the-latest-chatgpt-claude-and-gemini/index.md diff --git a/content/posts/2024-07-08-exploring-an-ai-product-idea-with-the-latest-chatgpt-claude-and-gemini/chatbot-arena-leaderboard-2024-07-08.webp b/content/posts/2024-07-08-exploring-an-ai-product-idea-with-the-latest-chatgpt-claude-and-gemini/chatbot-arena-leaderboard-2024-07-08.webp new file mode 100644 index 0000000000000000000000000000000000000000..ffd2cdce55eee56d4a5823c47d5253d03198c11d GIT binary patch literal 37692 zcmaHSb9kgpv-iZdolUag#v9wl#lYt{qCxs83hSZ(eDWWfU1a~yox-#8Y}<+p!~dQzyVI600|)>`CQP?w*b)GpN3Wr z;0yqOm9?Y2f|ww&nz{xt)G+`G0R6WcK&)@*U@IUmFa1~ge{YY=0Fb{u3$%a7`rl># zPcDLyv4bH107CpJ!Eb15@AygUebTHhj<$d4)K41S&`jUxldk-vzuA9I@RMHs%QyHJ zeg2m={fGW%QRwzc3PPWgNB`s#oBj{l;D6ADX7*N}GT%RC$c(J4KgS1E`xkBWm-hTi zTm5wUoZCO`?+W3Kt(6r&?_{4BK0pj00gwU61Bd|z04IPs;3vQlK>K;O{^U3S6h8X} z{||lKfAr-)wG2MB%mId<8o~e@fE7UhAAP`I8+^)qw*Ty{g9$U!KP(Vz?^8wkAn z3IITT+W0*d2>g==1it430AQN{K(F;LmPV4ts`CME#DZ3zH?pbh|_%>n?hn*Z9j&%VF?fPxhOK0Ans%=L1Cvk?Nn3k|+9SzaHL(p5trgO`n?=w0qs$mpMl zBlyR{QQN8FpOp6;dEg{)_T%hNqff(o!;{2&%m>ur%qY}j&MhAh_~^0gbNxsR!~?1V zfgeP^z~h|XKEOW`K9}$3&+0%3V9wIY$H-@w&Mpz~y`{(32YAMIGV^$KzSEqu;{DCl z@5Sdbr-j$^L+pO%7hfYV4Vd))XM^b71|iD^m@y0d4YUU;zb`y#E+f6^ZuO9OeSKc~ zpm^B1??J2O+5EzZtQiuOfBEfdoL)W40@kAcvCZXuDC_0XAmAFIkMBY*R+G9{J|_6zgQ{}S z8{=u&XJbzCYVgva&)H%W8BFo78yrtFti)g9^ir1J&I-turz=)W z?0$&DqmSg=+t60=oW;F-Uk~~Ujl=o0tV9!-H z+Gf)UD`+#pUJZ+-RJB?{r~hr!HdlZj4y%m5`aa$8fwzGj3}y>55W?f|s?PYwCluq) z$~Rn9$QV0mkcJR(&-2OVGl*tjitn5Dh&LU#lRL@qqBG`K0t4(8+%=8`cawB8CxtWP zEAG<7X+^%h{W2gs{A7&}z%GFWR0+uZu*#2&PTn7X9=*Yoa*;+{=i(Mt*m8L{)Mq?5 zR>Kp_S7W?y;VDtrwZeCUZUfgzOGK%+TB; zg6kfFtVGP*#=!>v_e39mA9G**YWN~WC_dU z8XNC?b>2Y`?PS5l!Pa78f`yaimKgD^;c?Y9dW$iIhjj4cDfp!6KHTUm%b91_eIJc| z+5QPpl?e{h>(dwP_38EKU)6?&0`mnI`(l3khuWF?#^m$G1KqQ(KxReNI*h0YtAE-QAylTNFJcg1Wde<7yd;>t_F+R&rmw{v|ux+M4KS9=op7}!L@Irdq*4#b81 zT^syqIWJ4oodrT2Pl3nS@L7~emz88-0mRrC(EN9S48hUh@}I&XUQP znwFW|;`9X%Zd|B`ce==#-LeB!G%sg=x%b9gwaxp6^ zz^kSZSxswUY zE@D>`VdjS0E;nA)7X2ain2Apq(ANzq@Fzbu4RVfpc=$T*3BhC(#QoDi7fLv>E&Q6^ z`YY%{Re z^j45LUKnP;^l5k~OqZ4Pu!_SfNCrSM@_d-`BQy8**&A)j_I?l#dqIp^ZkB>*X4|86 z6FAj6HdRa%wZ$mItnB2FKG5FJ&PgY-f_D{CsOj~x4F@*ar*-z^^$f`%AdTzJ4LeYG zdBa{s7ZV#9Ow(eR+3_t?Q3E?)aN!V&AA??NSgbnE=PLmIF#2hp$Bh&3)hXqP9MbJf z2dxL3IGQHUxSL^PRfNV{nK$4QzpgktfJzwbn(xwp+@r3QCv4_N=^oKhO*JdTkQ~dD zlD4m(xayvB&fo7293k-hBjT#q^~e6%js|1bH6heQ1bKf(k+k(s@?)DTC~BEyA5ZH^ z7wu2089Er{m&rkivHzG6g@Ua7zaC%zGJXQ|zm0D%S>pjWhUNoC-UR(vSqo(HAr z&_%&!)t6aSVOobW7OhVGs?%Ulm(D|Za@okumM8Q+-e~lV@s0LA^Az3|Oim1tyA;Gy z8_PavcIFPN+>INPd?=ec6iYYMFxR~Uy%TznvEenRCR ztu8*}B|A&2QmXPlA)|P9i!W(L=As%$yEi%WQBILyKVkfIP$q4#F(!)SB^yr6xN96NmbwJO=%h?jW@$WGmC9U3hglI_BFyb{0gTo;NvLEXEFx_=N{##bMc$ByhOT7-zP`U z2z;iNPh+LMm)_nm^ND-Rq~DV0fUS)!Wy@a8-`fHn2tET*v|Ho=(umM+a8HBs7K-y* zu6L*iy|?EzP~JE!vT3KYiekj93vJAZgLKJ|7%sKKyGN8|fbkQiokH?1BN!o2NPXp= za#9jtpq5h2u_Y28+Onn+#7%;aEhZ^LFx~kfJrb9Kj&CW24J=19XI^ZgMJpuU1*#6H zc#b3p(qYYSrp)7>-U7Jr)jtCx#b;A}S7LO%<1>@IRwtFyi`y=<(}B3!D*u~Rp}@)E zptu-^(R9ZFW0Ff+`zf$6&P7aTv&Hl2*0QE6zldTs@=0y}dmufC&%C%}Dz z=YP$1i=tMO_cM8HZERLva{J7sD+>ZZGNscxMKkQQeyZ4`@Uo5K zB>kKPm36A#Tf!NW3%;Ob^XDf8XhMdSot==CHOyiGon*N z;^^z%XO8xnf1mc0wcaO?`KJlB$WHuy%DE=Ft~2_n){~+D1dJSWQd!7wWM$kywMb8eKbLZ$((}G1)RLH*pv(U!98gZi9uWB6jLGHRoNe&g*GkGl^^;5zn=ev$8qCsB6 zWzBZM4I65;7$GXD#O-_Z8l$TQ3sh7uZ`PECrexj)FT4@N#|L%(C!BSnV4{Nn(w!GW zo2!1aYH1Kcn*a6h9%gfGGd4LZ`}qiA>F!}^8Aj-RtQFT`XF^d>i)+j55ezO3mE3^fr9@Y<*17L{CTKF zg^AcTb3G$mF~$A*1E6Dlo-YAVZGBnBk$2j$xC>3uB>WKYsH9cL^ZF;t6xOX7i9s^P ztaqkzW@!cR!^Vqv&)~on_^99OxhGh_7Nw&snlVa?!puLIL<*Y7N>zv~xui0;1o#sy zs7|(R=OnTpvp>#7+w6G@AENYK=*QmlGE6aW#OP78LK{pba{C}yjKkE+b8EhsHgY!f zc@S?VUnemT@K2Yz=b+Z0e#@(~xLVSj8i(P8vP+J%ysn{Wzd8Co8n${};dhZOvM(<| z)4Ns5J2R;{R|;Y-C&dsMsl~Fxk{CW+dUc_EKq^P*Q`U0ldpT@hd6P&GZTIq+wUmw_ zeF|tS5=Wtwh6(hQJ=hV$<pSFl0}cS9+tsaov45glik>W>SZxeD!b({xxGYWX}_15>JF7tx!g z>7LTRQzgP>8HL_3L2oteIW~*AChFvN|PX6kb&0LQ#ZeTS}VHImdr z6lHaU2yUJATBNWXhp{UfllRiSP9Gy>9?PCUXw?eRYmpwVqLxVg)or0Vltlzt$7_6% zFO>U;p+&&)7|~-DCTnPoLebzJM_A;KB-ex1kPj{&^qY8P=k+y=9P9xx$Bgg!SS=<8 zjKUeMnB?%YevNgr-Mdf$NsY&$)C0TZz^(qdzQ2;i{xNoj+0_O7a+Mzgj16fFWKit_ zUcZJ1bS8Yk2+MhNeHU~ltg~G0(`Yln%z3v9H-V_D)O&y!df!LcuE%#IT}-$(!-Zbr zs?xVt0?VScSYVqKy8Q}?%Mub1?4dmM$ATtX@_24oiU!(hban)G*+|(X#EoOiE(*<< z&Gbz`p$(jCc|2Jdf51HXfOm}#94Ma(Yi5i6?-v(~l;QFQl~k(4j0mp-MR+1-cxjLb zKm%u#D)BU>V*VnQW;}U=?C$})l^Pm#RNoAz$7v%>Y^czxjB~T9U8O3%bMP^I#Dy;B z?lpvh60=4~JXM&H^6$Oovx#yECXrw17HPfuAVN$ZGAFy;!1HP!$z|WmC|()u(bV_H zzp3aP3w(^+Kx@9awWp?C)!op6X2I@=3-jxjUYCMs%*tR}u&@M9g1j>nj6bA_fy&a> z=?RgAht9VWCUumDsReyqw@^KdCBe<9e_h)mOP&4j(LYjkKYN@?#Qw3;>FxjhcetSE zi^TZsILI3&_VUQglTma@)SBq8!s_-K^S7OliLr8pZD@}#s7><_(fDqGOir=9K4$Er zyl9y~+hzn)YOmJ)z#1-mK5_RWYpY8uS;1cW^wJ&j#M{Lp#Rn9;9IC)c@>7|${;LAa zzB+B-`*RmzDP+Y(9Dn%r?hv}RGp`$&{qA?B4U&iP#MNAWgT%*0zj@PSsaw(ODx1!kgYQHiukq3K8Yw8|Ue!Qrjf8t?J+aBKiA4V>OxfH}cNV7r zOgD2*F|8+6HJjcnm+w1n5k+>Yc}x)JHA~oO%hFf%s&ozUJPz+_{Ql;cfai}r>grG; zEb|0mmJGo}$09=scWCD+K7$=po=1Yc@mdP*NUsL#*8#*zOqfrc)TqhIZ&5|MXQg?0 ztXcrjFPrXT5mG3*SOQ2{@`4#75(DBu>}qEux`tMt1w1u)zw38ymiOZwie8$o^7IGl zipK`kr1duUlMT%z47N{&jAZfNTXvyg^Xl0lV^6QU&8B2PYCoyh#9Gytv%4^z_b=60 zt0%2i{`M5sh+@q9RcMlpPRZoj>P>05^pztfAktV3Gh74WXWcBMgGIFxG20_pp(~Ao zTNiQMJB;$>;IIz`#42EUGTfY{x&T_d8pE#zZXF66KaGWjXG#sMfJkuSy9c&3gc@kG zt4}iuuQq4~DREq~?rPjyYLBCdst^@s=G89k3V|)C>Bz4k^rtkcTNXVzYAQmMhX)eQ zm+7@xIbkvkvlC8NTgj_1SLL~7kuSSpnWA-6;POPz4GD~^o&8oIKjN|Rp>)L$rk}B{ zc3Vvx%{e38Q%40=x>O1#JaIEh=s_OaxcPo7mfQy~&20GikG88WUDc;aHR@@#s20-n z9ln6P+gcFS=Q;TSjkHsWXpeB?wf`*p!Kf?iTKvK=gSPXoO2_HEHs>I*{{?|4*a zeMgVYV)am|o>HO=7qT^DJBZVevTEHYUw=zXfvS$)*L1Un5*sFIEL3*sTq%T|dd zvTWoF58RB@3epWg**-29app3wyNzBBQ;*=&*GtMQ=zXXOP_gptAnwGGX~<4B+eO#w`?34(!D5Ubkm`GvY2g$}r?o08!c!o+R_o6lfY!Ua@OV(B+bz zrjiU9FekuyB}8`5 zsAak}6Xi?{ZfCHt!*U2LH9aN25r3K9h%=z*k%ykX3qTF-pQ(GUI3(vQG`=2k-!Qkg z2Zl?HTPQrTN8GM9TvLwrczAo($c|mP*p1m&>MoSVf_3!;rZ%xmT~|oIFs!KO+Z?Il1l5gM z3s{`gNhAH7u&k)hPGi&sH4cX!;y_%W4x)A96@SifJjOf4#0x$UaZDC4kb-(rgc52p z>^FZ+!kC;(GN^{4*e~fVW3Etj=(fRz3giGY;O6S@nD51O=esaOq7beyv#@4|(MX^)@Q>c@V!qy`>vjPGtfKG;>9Wq-OBO=*)8^q$b z{0)j9(vJ+&yUi5pZ8&uQD(^x9Ikgxc|Hgcgw^y@}2qMRHjV#ShL;JgOJi{*T?cJB^>x2YZia^bYf)?AdO2%WE>=}y> znEjx{Bdbl)h|=>IYh@<(=)g?WoU9j{EYvg&F7uxn^192dr51O1 zJ9+w%aiS(ga>;P5hD-eri!?JLC!KgmVov_z@Wbi92x7Boz06OrJQ`!C`KU*0pR%y4 zr5!{D3-Kp@&j$jU?E(T_LY_w_id_~MDeNP%oCi~f(nHLOl3tgde>9Q14QwALqDacs7yT*vN`52TVS z^TW10=1~}Ws5jvzXPW(J2P?YN{BW~O9xX{~Ou%MzhdPecRGwu?coW+-^Tpb3%S}7) zo3vJMCU>Plt|z-fHcb;1Z5DD`&uR4S>WGE)?J53e$n(7(pJo(j$TVxKsYiX zCg8|hjfn}4A8z_xxGBi4^CQ-e7S3p^8GtuhH2s7@;$Y0#0t*E2p@&lFN}%2R^VQpwI?PUFP--b zu7U>ph*Q+5{at*3*UH#PU~-OPcpthVxFp;w24)IbP4RW$&y$F>3-S(Ax({|zlW)$c zc9c!+(etg>Xs)jn_(gcz$k{ZawYZ`<`qWufj~QJ>c<=kRAzgLlH0q7|T4&^21=y75 zH1erHHgjn0Q}x zCA(b~0eqd(PF~y(DJ(aq9odbSwSTpB(EQYur&J{&Ek|)G4l7Livlu-5s#?)L?+YsD zf}2|z{El!egW6XTldffyj&CF2<};s(2mJX1F(}6|K<+#GANys-RC}p@2ea)ZbCJcu zfNFW&)R!bgvEdmRS`>U*sDX20Elj1UGUj3cY7TFS_Tsm#JH_2SgAv6<`rjSGg~4== z^!-G(5jECFBzvJ(^d!SSF=<<+=kEB6VLif@WavhMz-ZuD4?5! z&|4oq66BlYFw?zB1uvH=bRg^3N7Sy;!z;7?wqrQ$H`dn=@zd&ndW1CbQeDmLRd5JE zj8R+z!62Hnd#7g4QQR=kGNjO95^(m?e4G1+5$tbr!7!-Tz0&W;hgjNKb&T9T+r_g* zY+Z7g-O(n)ZI*2}djlcH?%cUWzH{8}xN5C+1=8Ed1l9O=Q-EH1K!z&|Kkwjiwpl#iV3>BA9f18>%U48Yx=KK!o ze=>0!7KfFsjL3Ov8w9-(CqND~#WWDU=v2Ab)TK>#c}~JQcMr8;op}x4e4k?`z+#iL zwQGfs9(>aw`vQLF>C@r!H2ISQ5d>tTTW@?OlJ&TC#4 z5&bNhAlLbEr3s8VQfpf@f-SPvDB2`?)q&f0gn4zGC{^@Ek*DBNH(aJ_q}O=1?Cr~~ zq%DdB#^A|I5x&VEJLmm6y(0ZpH2YQ{9V9spdU)PP9Fm1M?(rVFssZJ23tOr|kdyA} zkJIJ$`5&f4=n;BZxlHX8Vd&`+>;&i6gUyN~*K@zS`kE ztU!U&?q1HLfIzbcS}I9lTtX_xR>hBv(6wW{5KnCo=n2*B2)ay2X!eoIw9b{C476>n z6R&9L_=iyoa(F)f6r92`Mm$!hZ&2ocworm1Sj|Yk2+n2L4op?k8Z}}aPF;E7r1!u+ z+;$HUBG7^?G&$(KYRQD{m92lHUW8&6IgB5FbE@@?JRwGHrV}ajb=reWC?JgHqT%s0FN;M75#YLCTcsCX&a+DX$PP&zMm+7d~b%68&_#!*9A zz6VAN${!;tHFt7Qoi#8kL?s^XvUu%WB8lDrboYZojF8M~8G{ew z?x*hU-A0j2&_1M$E@^YjC|O@{ym-(FNW>7~UkN`sNF{KDjfSUkGj_os@wKadznPmU zL+KZ)Fj0&DHimcUnvroJ!Ld{Pg$nl#;>OyPyk@jLv^^{RHwuGZdJduxVq>>_iq;Q)8DL=Y6sx7-~{mU)~HsHI0SIJPZ&`kqCw~a=@WSuzB zK6*O98N5NP#)4FmiPf9dJ)`mM!Y|*khIK)tp;Uh%TxH{TbRw}O8w_?V((ihzRQwTP z6r^ytyJWAY$LxEnB#TUNWgxelOZH<&mMykfbVohGr{@z`9J_IQQ|IE)t}fQ(>jSx& zYTuydLW4bT8QvFyx1y)sVSHbI&PB6r=eU8ZTq!aqsg(q9GF z7iWkF)r#C``JkvvJ+$f1(M<&BeS<1YsE+H#W{$bwh@vt|vcYE0fYe}3bMU?Wc&aI* zMneQbty;htq~L%iWrm<%h-jWfT!_z#_5h(*JfQ{+W>P{*h7bKeyCG-3CZ2(9Gk+}<{xcr2c% zzO>$|K_jLBeZzE5eW?x@V(^E4F%kW+xTbtt36wpDXyKMGZl{i=i5lTwBhO{OGsbgJ zo{tC>kD%N7hfx^7#PT=$Xt~+8qBS`0AN=ACCpHZMLhn5{1^$SKt2z2{$YETjZwQgd zkv@q%LPpJ2IPTp7eXz*3%`tO=lVUIy6$;^HPuza80hfkNq9$B3&iAm{njkk-Lmfqaqv35?p)wXL(Qk!LA-_yDu)`o z)EFf|FxEC;je5f8N!gAdEaFA!g@w@p+Gz`$fCxq1EiZ`HPO4A)R?5J&OyfuZ_j~M3 z`bZ%G;o`DI`;MG(m;b^z(6nvZ1OQsYSUr?`)VK-hHt}B$Z^;^uReD-fvp$0WwJ46;)ghlhd){Qhoq>8a&I*Z$m{MB zx!0g3je>ge4l;az1wP8o(p^ss6R&upv*XBQ3iuOzac$GqomqHkZ@{sJRz@70BR<77 zqQ$YY?Zvi2Szlx0R2&@!VSm9jdDG*4QpeU11H!|!DRuQ)lIeYq%ppX~G*^0F&NzO; zy+aS4VM3-Z?nh+wuXp_%|`LPBhn!vTYP%LwIzIER1%npG+|KO>15s zLfeZCdKnD<{sC00=r|G#7cIgNEWVBYH!~Ex&Vu~v^WvMSv>{crE`t{x#C4Zw_b?27 zn`jum4P#U6hZzNaNLe)H$-hO|HenIN;sqDBD(#$xg* ze3)(=cl97n?;!A<<4McUk0h%|m&Xw9IVUhCIy>wlZN;ka?9GLSRY@^gRdRNbtgGza z_Fcvk5RUMLEC5-Q6=Dj!|5x!GyY{ApRJWRg-1N z-G-w>ESnlS?2hFngVmc{ks%|~_rliHmaCX60|UrxV_B}iI`p%20+B*1nCD8^htVW?#L0m1L+iQ$u0Wd2~Z=bq~h5x4O_;cQ3(AHy;$Yc zd+`%P=YEnyxo?EwDfITMT~Zl7lEGtf$};fvxpG5cpi6~Y@ld2>SKJ+iP24CjDWK8Z8)WUFe% zc_~eW+4R^}g@W}LaJGW;AZ5;XUth*F=e^P4r@ji;e&_CY>F4C=>ytnJ#?_kD#N5fJ za&Vv!i3IUP`CWK*bAHZOotcDFjoScjj)tLVaLO3{I&BNQ#-&h4I0C zpu%>h`79$cYe~=q5Usb;HiD#dP{Pe$z75d|L@#0hAxiU=p+hj4CWIhRiPPNpno)$H z%e5MzN3LGa(C-ymTVwO?9OG(#F#F|05eHbGLa@-E+PJb9^e6(+eJN!%mq8E=I%|l1 zbx6U6Ks~|#O1ZG4^XnZQ_eVU}v~L`T1`y@BEN2ux!PGE76}PBqJ!h?i zsfvjqYat@h@&><6e(M5o}y38;9hY-ge9FXR$_2b3C99|Y)Al*uj zt_Q>aVjw`_xv|4t7*PC6Dz1akkfGOvw|)7$p)*7T@= zp;|oMh}|-1k@(|M0Td%KN=qptCemhyK#{t&Wl!utM--9Vj_Z(iA7OrJ(i6>ax}qf> z)_B0WbCxEP4mq%JP%_@qyVl{d%8)-XM!smb-!IHpb4+aU2%;ANAP9cbW-hr>-bG<+hm{zJpGjFp&Y+j(E50; zG|UiKyPTQFhWCUWMgo_CNwUW^^5?`5;lirYA%IYtDaxq&+@3QcSLR!@IZfEJw}iKV z2^s{riTR2`RBd>Wz&gWq-)jU=r!p=x6{SaP>k#KOnJMb@*79etU)iLxn^|hz(vv|u zt^ODv-rtKTaZWw37P2H~K}znl0y@%Zb3{Er)-U)Ki1LkQBS;qeks)(^jn+0;EQiS; zzk|yl6Qw*+P>sBDaapVB6?JX4yB3WP5H)QaMK&C$y$e$&kJoNQ*sk)qk0ZS3rAaHyB}!7s3|QUN}(TB93$Xw4+dcRdF+oO!MpQ@6=+PW;N1#Apl|k};)GmT z=7Y9G&&0*6%{ z7j;{|R6p>`L}yx&C%?QT&~T|YaF@~lag5^&^0wE(XaQCTO>?@; zK^_Z46G$lRDKkgTJ|TTwL>CGT&Ne03581!s-XDEES3Np3L5u4-a|hHSp)w~cF;y3> zdIi&?C#e_%~xsUa@tYe>3Ey3DwHU;*JLlb z?oxe-QT-lJdhapzOrN9GRa#nfV#h6&QO4Sf2hH4cz@|DTCj)a7oHt!DWp|5WI14M> zNrmake~VbtIDS*2$I zNw$Zk7SB3RF)AMvxR97%ES|~D#J$7!^SNyB%M^s@wgR4!1Xg9t+d$W^(G`?eX-EVz z&kHuaN~=qm8Y7(TyjB5hlZ^8Z7~aC4Y^Q(BXriowe|M?xWK?#E1A$q37-`v#jJ|l# zMb>G!t7Ozj^#zcjgSz=yNW1Mwd9b2!QX5#K<9MpTSWel;7$6kFv;p@#d?04#JxQmA`GBXMZUbUg+Yurj}uHO`FYk$HGDTMdD;Ss zxW6haW)%-S?ZWXy&C5as3Eq(kCEgGUk|aM2cV%Jd-O6JX>NgbZOf)a}5^_-n50-27 zyhwWf@T2@J4Tlh_=fdL=SS#N-ZqqJpg?%$i#q})nre)d;HCQ}QmiLRZR@%4_t8|Z? zE|znHM?@>Yk5pFYQoff-9_QUSskR5NFki%5#38NPXlt7r;r6g7YGnPn;E9=y;CTMT z3^&qGxvV3u#=fxmyQ9cc1kdXJix=J~%A;Bj<@cx?@J_wmc8(;+!tMA>qA__=%UGho z0X@;1@*c+tmxS9Wxo@S(holI!)I({?_wLSE^i5iZD6zt#HMjKl986&m55^LDbteSgyq0EVotav&RI(abEZ{!+; zAI*XzFeJ(q?0Vow2AXkX$e$;ogezu0A*<~)^}tPn=lmW#`1J#^h5!IKXB#~@a)Q-6 zw_FG;%CVLvon`bWVxn|@T%3feaqM~9@avkL;0n)Qk5W5BZ%mLOphqCSXyfna805M| zVaEumgI!HzbU01)&r$q=@G8abW}($}LEuwTaLmLYvx6EQwO!FqQ@ODcw2!j+J;6b0 zf}o=~O(pcs)cyA?bOsAmL6*|f$Y9GR%~@`qKQEIn(qz}cD;6>E{6_ne4AJNyaXoTZ zCnhreX1jD(u5N!k5p-g1?10{RB*Cg^5GGy8=NKtqHCr+jaIZNF_c_js!qiZV_L>?0 zz)bh@PkxeqY9=N@($LLyt2u5k+gkJX+God+NylUMZq$zh!Y)@{3k8NZq8E4(*oc0Y zYe|FGLY5VmiB{Du^DFEH>m{6q{sYR6vStzh{jsuCI#A?r>-DyO+)s!Evj=IQGZ%E{ z|KlyuWA!WAHB8v}>Ch5eXT%i@$B>|!P*eTX=p^!%@%=lli;k<~Wu4%jQ{%wKJyKS? zAD=3Ng38pJ66S9r$sGI$^d--Ln?`k-s))k;OUUP!G7MPD!U-lWv2p?WNg7*H99h0~ zsI^IuHD%&`9@M)rfoEb}IbuW=PQ72SspltQ8IA!UzDwo(W9Pyq)hOJX=vxw7%>=mI z;3ZpUB%?xl{gT@XQ6jxeIBxKYnk5#X3!F{h?R$wKvA?a9_4?WR&!2jVUf6a~9Rq~Y zt?Kq~Srlm0f090tk9ZzerPYzxsLC+?0ef6l!Bz{~+T1hTSsxFzQZ8N zx>-0Yg?RPoI`&Uh7M>ghxR(8@5MwK9gs}^R-XOg6UJlWYXs}KA5w#`bLs^AYWW0xB zIcbn>SBM)fWz|`Y zb^zlE;4dFkFsthDG9RNy#Dzl4OqGHwQ(Lp3)hj^Jw|3vDU1vK6)9Q*JA*A{Z3w-Vk z3E5ao=3$?DmbU?|m~4%cY)Sj0vE_X4%)~hZ?y1PJI&#{WqY^AeA@|3XHDiE?1D&wh zQbKWl_0_?4pU>82vaQR z7pIZfXnf%u;ac5JBt@eG_u9>9ya|&|xp*1c?O9v6i7tWDdGmUNJqL(QF(U|LBwOTq z*F>zphI~cg7Rhh-yX-6!;l|(zfyCGiKJ_ z@ivV@9a4rHnUt@x_z(E!Y-i$6uV*#&-1Eb(k{rH1PPvmOWqZFXTFO#pe4rjgdReji zpL|6```T9XfepI#ZfBr#GayyL&Yog*+#)|(b;SS%u3QbLxbD+(zlo?{{X`Ov#reJ- zZxaf7Fv)vE%W+qCbIK3P^qZBpyxU{uup)1w4W~Gnp@;FZ#@11Tb4#Da+{+L5Im4f}4FL{fo`=P4?tsA;YL-W$< za(JD5V^CG>t*WBuk5|*3`!AA#*Uo@nsI;w1K%1?rN$!f%j>+VOPh{0e8($q0ePn?t_x;STQbFTG%n+q|dzlY_lJ6;noVa?c1HOVdW1+Pyt@NUR%Pv$7Y)03_2b(R$~0;_*7V?bb6h0(dH z_^(VwPqU{X=~Nr4X?DE_H7L#(WnrAasTS)`l>WXq8INe;5O3`}oui~=@FoBkjz|z5qV9XMIF@|cY7)z9kI)6TQR78G z^D_@G4}|~2(5yQIB32vXi;ycxCa+^OpwCOXDu6MWMV;z-_p)elyN;^Qc06l*E)+Y; zBR2nqGE8Al6rw`eU!AbIE=RJD(@F6F9;II_H7oj*N0`lYElK@7ykr#A9LOgvrIFg4ze|;m)sl_%i)H(V)nt;oL^5WhkkSp7aH+lQ zP!K{uZIszW>~AHvWLh7x<23DlNB@lF>nNfzMDswg#DGy9Kv4j=$9|**#4#67at@m6 zxj!A<^u1*euCi*K=lXbHiMjejx?i7dQiE0|sC8m>_Cqi8C26bc`aVxk-99JFnA^SE zGeT0mg$Yh{u;Rg1K)4$ielX+|0zyMu6;TE`W^m~aa{}W^PAk*A=0)Po z2vmJXzr_AtlAkGadYx{$LRw0NRHca`rTA3%+$SP6%GVO{g;kj{!nCLItV%W0dS%cQMM!=3*$ zDL!@m+Rw$N7{Rqb0;S8KVW=w=3GM^VmJ#I;^#N0VTwzV|L-B;(nSW9$72Soc*P9{{ zVQ*J*N0otJs~vA$dHSb};$Gyd+NI3cHbDZL&SCxV`E8%GmRjk`!Fs(S%D!<4-B^+= zGG{z#s*xngJl5}4eDkF1N+-dO_!G~gClBs(Sb-8ki(NuSFfH@Z;tk@yOo^o)kbDWu zq6PRWd%bDzJp6SF0lZ8dqMo9m^~h{Fx%8~(B1B<}WMok%JvPpxdAYB6Ke~AY#xt6h zyL-e|24jD?2+xSk5fWmrL#^m*LzIrudDUWE>TZPi@yIGIZ!nP6C@aI!W)~zGLz1uz>bD+sD&l(hVvwMShmGoMGc3cbrVQ98$Q<2MtTR)yY$4c@us^C zfcA?M@cP8B8$`Z>q;u^0SAy7BqqVM~wr~6jDIS8f&*ZhJ#azhvDnD$p6s9T0Z)}MVZ&@on`!djp!u1O^G@3MlCW+UEM zgFg%E(Q1~Gn!3b&JBhhkK+wk`myHrnc=KDMeQDKT{Z`2lfDH@C>$ov1HllkTlN+*=qoF{S1IeN55BQ+ZW{(g!Cp=Wz&Ouwyj4iwfsQ=j=74Imc~jYi&J+U z?b?4-4L{=mZ=E79LqP(%SO@}e!vK2Tsi<(_5@X1T>gX4Y34_`_1Ly+EG0T%!={5U{ zu58*og!jeR?p!kTG>;Xe^FZgl$p}G^a*oSEL_c>1C=)?Z3feZY_SM-iiNSLFIY4cC z`&PG*?-|>97OL1bJx+03rBW$Vwjp?l_1 zC()={vOopfL(Tx1#K|7EOb##0TU5bk#t17QmE4>JT#{(Pwjm3)0V!eRs!3Kx?3TwnQ7GV8Unk^`l((oO0wOPSr|vpbjoNCG z{8rH1V+&`%ti>_ZxJQNQW*0hl$sbfIM{!LH7TiM(8z>eP>OUCH$jKjykZ`^@`qMwS zB>Zf&Fs;GZ(id-Q(p?Dm7|uQ*vxrN%0cWhn>Ai8BKD+hOVe?jav%12s@aV`Zr?Va{ zAIM#<{DbG~jrHqSN?csmH#&3bL@W{sL15&q7BVATP$LF>7X z6wERgmQ~vdcEW2}4t&RaGZ-t-Tq;(ICnhRT^1vt(1L=)UE*o`FF43#-?M~32`JkJ- zGj2VI#Ep+@1B(6Ge|{y#Rt48)1yS%^EoniqMh^Z*GAzuY;A*-cHvNu?!8EgM!Oa^| zm_ll>y78`AToqU_^4!GLBKNG62?A}2iMeUlyi+=iRAUMi2$5ZbnKR*D9?*W~E*P-d@8}^owFJ8rjo4Mq{BsX2s~O zKEstOcF{-LFshT1SvXjB?%o*EL*RSRR^QSjl+3{;letwJOxJI%*qY+VeAa3aG8r&C z1qR*dQS=MXnpBoc#YS6kK^3^|CwNJ`J5_Gh3aS>3jfg>)2v`%&rEM(_YdLRbQYcR4 z<3ZHIld?jpa2=EeZ6yClzoy5!hYz@i{Kx?4q1ANH=2n-(gfJh2%#^)=x^k_2y*H6< zaZ1YJCrdW5z`|-;-4B_;=XIno&*^)rK+j=M$tS}BCwDC3&W~oSAt54?=?4G(4@$Qj zRcwMUGxW{vKNU!Qq4}6r6L9iOpR^OjituWGOS*Bo%5ka`Aee;Qy(B>fO`n7oBu9jm zPw&_kXE*8VYP}`V53H#QcbmYW`6{b@K6iUl!CfV?;rZA^pU;6X?w*5k`3|F#7aoq3 z+{wAr6=*L>5aCbG%z12;|30O>pRXY1TA3($*2lz0PwA{wQTUkI%qR_s)9kH2!~1hj z|AF;`Q2|DRCh5wNf9G!39r&fnJsB(vRK)b=Q39k!!uMynZ`L+~^{3 zm~G&m;BB8e&RwY@gQUrEwGfLE9X|xC%UbN?jh>C1RaB}hI36GFPsY7AwpJ=$4v?H-Yt9Fcje_c+pFnjpkQH@Pv67RG;@Si$3uRV;yIGl&M zFOb=4ZmmY@>m|j%O;zXwZNFXQLE52uye)O4n(k(xcl%Qa%HUe&WMO_ z5HWHErUrj*u8>zvHd0`^um0b?UcRF*okNKtbdF z{Aqsn!*tprckius`(%9aHa~QSPnKHogsXXJHDifT&P^iGD<$wz34%HBp;I1dOG$~t zhmqQ0o)GIy#PX_Af_!;>!n1)-jn;k(P(X?3i3@qZ%S|AUSVdiH>t8qQ<`??6_8`wW zNOt_YB1wwuxrcjnDG+j!67h*mmm3bh@1rVpkUs1IvLT^Wpyt*=)iu<{IINiH{E zxRI~rs=84}Nu!qpp71)pHK&rW7~$6>WKLxLwRN`5RW>VlU+hM5%YzueHuY!!m`s+| zh0Mm9$!f$^iYD>%$DDmY(W?!9% zzAgohFr>^t;Iic9n^FEZ8)JibC-|^k{oL400d%MZ_7@L-2L}d}Vi$K0lGvGStdP-x zYc(Q<7sgzknl&G=M%x)KF|D?G8ZAj?a>d{3g(A^R&zen*5)xJq5!AKW-gn_T9`3n* zBVaDBeD2&9JXp40Cy?_;QC%{xNeF~zN{GLg6NcH&1d#;0avS&vM{(GPs1{^U(5`9` zqwcs;fhKT?Jvx{~_9$U}wUkWgVnlDtEK)0;d2RYU!Hic+oy4~Ln)JTkV>==-MZfnU zV#?e_vgK=uc)8vUEKhoyt7sMoYFDD}hb1xSTF`J%DWglFqm>Q3-Xy9-n1_Ya9!zDj ztO2M%c|wgvV<6c%mo37#M=hsYR`~q**_$a{frRS_WArWoM}Z=$=VzL>G~cN*j4_)N ztE;+Sy?Y`4L=h&=(AruFm3TUG_o|+yy1Uv!Q>ZvH~0Q+mr6kQB}t@wq#otrG6^w&1cbL5yR&he6WUE%coJhY6^-UVD6V<-3GKd_f-Xx5DP8J zRlLimZDF`oJyA4GhpOMJN3VmrO$qSy)$ek?i&zIDQ0pcGULoBrxS?>iX4QEb66VzSJL?`r}A>r%!^5ghO=9kX-}cmIKc z99x8>p9JY7#Q4J2cp4!$+9^BU)geO|k&58*0_2cb75H!oI_=64rhneM>? zl`F(*z_kpDh!!?Wm#D`q{#zh&8QzBf;G5`sK1*QK9vf^Dyee%^ zGY!Gdd2aZ|zGJp|_?AAdF4FQ$aDFNVS>IoXD(H~k3gK<+W$9ORC;4oXhHAt53Vzu# zvN=%HJXc^g2)oND8y(?1L3vH3XN{)DbE%B& zYy&;rbCiu_F#MAnb)#moDw}y<88w9ZBRpo%S&WAM?-)~RybwFL?I8btsjVu&v>z-T zcQUjfByv<4NI-3CJ=>$`((}&-RtYG&pmX%P55nM~EG4W3?g%>6TUMC+LaNTr&xaZX zJ>dK(j#jWElqI9n{X$O~%cC$3WyZu^&s#I=q=~ib%zKVk(p0j_MM6oEt<3k-zA#bJ8Hlgv(%AQ0-hHQpZ2gm#c4nw7tgV(q4E%Dw}y!73n zH@3a+434GeY6P&~+X0(-VSt5smHb1*%Yo9zhEC~e7Z0N>I%lefl0ltTf`FishKv$l zx!YzsskhXZpEyUv^~@Ut;Tf+PxkOCx0K~D9Pg_X2@gfon;!E90^>Dk1+V-|`+`w%- z>`(H?+YXiPL(9-`OaRqwzCHQos{R48>RmGtO|qO1a-h6G4}AOrKnE9Mc;GwGT1=4a zlb(5{6wr9vaWZTOC`x;+u0K-^@OV>&2=x8?+2m_y4wcNyxBUZ%XYw2f3o$c>y+@Y_ z94ytpZvr}2I=b(1-Jco$6#!-;KtcoS z`2o?Z{0sc!%ML=QjnAX(`MXoxl$r-f?&2ym33@h-@3~6Z>B;J&T$;L6XH8Wc z$6`z(!o$hkCV?(vQua+6cIA9~PqM>y{q9DrdkVgOaXM(I!N+r_V?1^xmxlf}Oz<~^ z-ONeiY>$zhI<@MQ4|CV8EMNt-8fhKNnZ7Br%E608r7%XLmG z7SHu}iCKNW6BI5@qC+!YW1rT3{QanE_eHU_-7LPH!Xsac`Kn7W7-0nGQz&I{!n)AO zmSy4&9oPMr4=lqO1gJGri9g2#epkd0GA=De zB62fX8h%For#Sbx@a+2Tj}YeukaPz@9OWo#^ps3>!cunGuuN@SG3r0!U{?cwa33-A zTNJ9BpF<1P+z(PeoR^0w@`O*G?qSjPWoFM`oQ>-IO}77> zNj__bDOY^OF<^$)B~A#k4LJo0rA@}qsg7_GM4-5eVM2cDa3VNAr)k>}DugZ@0PU@W zNc_%J0AEP(lPb1dTvkuHkRT81vs5bL+I*N!5$!D$gv^x8TBTX^6Gh*o=KQ;Th#aE_ z5w#<7v~lL$-$F52T&Nghh6@H^51Q0S!f#NFfY;UD1R2dUbKs9O{o2syqq6L)u_~iW z^4u-JBq%d3xTD`p8JG)KxJV2;tDNS?69*MfDM(Ug-EGcWM?!e;2Ixvpg|PTq*q(n} zq$e^ar7uJZNL;)?*QZ`mi@v813uNsAd^D^z-z&;wD@DrdGt?lY5|vQI@CIEq+-46S47*7^{VylVJmBq zDv$Xhm)RlD3?_k#hkfCV0^EyIeh|qjF4*q43XAd&#(S_CehN~RTO(>VaWv-D;$_)= zYHWC5h0~qK)`8JD*dm8WWWIv3{$X5(v2gFs`>Zsjk!t<*!JEaJhOEkEypY1rBcf~@ zY**8NHtl=}zFqX@6n7P2193;Q@w=_7J;wYmJ8tRX2^)H7(@;ca-a*gdxYl|9177|M^Rs&du z$do+dX|bE`LKQN6puu+b%BFghxvSf*?f0AGLzn?lMq-vlA?!C#DsZX}tea`-^t?;4 zEY<95B@5tlO&jKttJpnb{md@u&b+@iUp45fK}fMXU0Edjp4xn%jOHzhlG?lTnk<(l zT`-8IYgP^?aOfZD^MqvXuW!FwSGTh3qO<3p#soP!V8RT-l4eQXie-ziS=)F{7_efHOUo=OFbdY@Lp0q+nR8lRZScY-Z^E1? zgF5^a)jP*e%eOl3qjo}ZxOu^ZDwwrMrVP&n!$=KDWUPTxc@a`&2O1yw1An`tk}Xya z7z-o0+?2{v-4Lkrv4(`Aq7ug<52?KMz{4qYMS$#AhS9M9wN8UoZK4D7Ng9lKkE`$` zh)q~ED4`Vsy`-eL=^L}wMk)8Memwpo;08nlbb49AikNcZOya(C?2 zJIUxgKTWIX*4GqEZZt)mJ% z7&=nunc|Ck>fL8FOhsl{j@z_*lmoe;qX@!v3Zti&kH z&PDj%s#qj}fsPEVw!rU0)u3l$Q`?}qZlVhO>qanOe6@enNBH(4sefk$6twSQiZYJJ zl^(Vy-1$r~3$-bE2`D)BR=J1=Pm|rAuF|aG*7e!2WQX;WF@PS)6dNtgCTp1@T{qq0 zq+d?uiOd~{n%k)HKvMVcoGC~Bc2CCApNl6Yy`g-uCz7f+O`N%ZLeb*rxURxmM93L#yj1q3BEQ`3YR&DTBQOEx^G&gou1XiV%?U_l+DK21_AO*?e6iwWitlM!Vu#Eo*eUS~(4G?hE%;nV82T|@ z7b#81b#SN$6jLCFfgUmY2%Wg9)-3kZZS)qep>&kJvJGCQgAHN%{dM(Bec(2eou6u;IrgN+_0~K(E(pTn*dj(z z3{iuElH@w6VfGKo{XSlYujzQRK$M|Vloz}_bU3>11#37~$O1idP!Hf`vJ`-0(A^5$F zzf2e3j7jKSNSqnNfG#QuNozeoSvD;<8-@r!A3v`_@>fZ!VA_W2&4KUjHK}t%Z|l)4 z%lkfg!A&Tq@9*%vX}1pp{8@E9UjaO`6hVZ<&`iH)*Os5#y;Rz+%U})##eY6xU2cW? z0)Qzgpk7+si}?K}D|5u}hplL`QzlG+SoQJ}VUTv9f-T&wp$@Mm!#2(HMduzxJvKrJFeh z44jtRBd0_gb@A7fR47P-P9qV)<-pVVe5LzcOoU~uG~O|fwP5QrUf;X(T);_|1el5+ z61q?IY9VAPAQ4f}dZMHig@hdla3w?YNQWnNKl~yQEKteB%Y=^!FaJF^R7ku9+~=|3 z54gUH#ultThQr#>oivTSh+z}LKMo$HQ&41o9v-Xb)@~=ZDACE3Y5z#f+17rTQB>>% zQ3rvPn^^abkzh1P4PFbjshpdH+=}0bE{JwsK}CAjq#iCid7*z7KpLL|=;ez3x3-OV$+Kcbj`~`BCqD8SuTJCtd(lrPUAZbb@Z75rlh9}sAGw2}h*s|O z{I6)6k@Fck?`;eNEz#veBCpnjHM$=QS%8!k(iQG!plsOYJ|&hJs~sw9gGmA(_BP6& z9GTANp(>#miN6Y?$(bjU`)o-+c_6@ag2ARQT;wGcnjFN3d3PcA3z*b(*6z&P4_V!W z!j)y6ol!ef6I`_C@THdmF5>c*yXUoSIJ3y?kGD6zkH?j}2t62FiR8~h+1dfggre0M zgSoZbb)>!eMn8?!uSau7%ATyOM|F$QOSJS;YY(7X=k4|5>;=>MJymhvnNz>vTXzy& z4hXZRrtI7ji-*SaM~3w&JPa<1i|Qw-Af*)qsmDvSJrVBd@~bl1qi%n>!Zt*76lC#@ zDFvc@pG^XQK^i+|f~hW?fWz(dww&_PwxdS~W=!tyNyw)cn!s$owUR zo}j&SvWA)*uY+JFGEUOWk|8-YGYNK(k0Yw}`O=5TJYhq$R{>pVqFQ%w8BHm-Xer!g z+uQoB{B>&1UG{bojVIoZyr>ylD?T3C-m_aI7~m$f&2MtuZP==U$gQbn`bSQMH5 zE(Tu5dPXmn-6-)*eSmdbjF$FJWq=JdR*O^XN~@@1o&4~o(p_>+vciDDC=Om=w^#YL>~o>CT_Sl4@~X=aFnvHaU?5%V30Ha&W3 zxScb;TLDU659wiE-ARle@RNLRa6V^Vd((9(^ch%<9WV2rK6I-_$eG}WXWQ_u+nCto zzn#FOxcPkta=wSu%4P9QK=OrH*GKa{#19{|G&jYt<{veqO*ABN)hQPZCOh~Ph(Ajv zzYj)|qyXs{r14WG=(o{@w8gTeicowXK$t60%V5&_us<&K&0QY?iALUo2t5Mtv|6ma zwmtnyBv`2#gscvTNgc>*4r~}wW7Jt*spk&eI%iJ%Oftj|gw$Fh{R*R%%MEN!YEv#n z8(79bOh1Koz82`$(<)vuU1#ia8DD%+pzFtvSmU!0Q*RpGd`pC7{3-a0x0lE^p4TTz@C(m~tsG-R{L` zM|vZL0sMe3XGxaC*uaLdb9TkfK>F;LkHJkjRZRKjqW8d(1q`y?x1Vjv^qC%JfORBR zTE0C3V_krEp%O&QxfF7>1{=t!&3n!zTeJLek!R5>!G&?P6*rcQZ`KTBH@8^9D!wm| zC6dyP|82pHp+6&78Rx+;8OIvKT|CQlxx1y@kKHA)*hp$7%liSMnoI+i3~qQ{sWlGg%*BX#^`8bzFKZZHT-~^GfrI*^wy8 z$z(5IEvM?v3DLbL^^IQ~hJ~2zt$B{z;BpdWRY=!c0+F?Wg+4Ch)YR=X#_2~f)R<}@C6_5ZkRU+vQvRxJ5i|}5T{Q{2 zi6{PPRxsh^FHoWQn>(XRnXW9zk6}Qjw-={UBpOChUR3E#)EvL5@GyTYL~Z+R+xkp0 z_pMLa)QjzUzSfLMd@CUWmZQR-$0N40a%Jb;-!t8U5n(tB$r?^F+(9(U-`0Ou8g1zCz8Za8{ z{&(D!@xOL@xc7g)U0qld2uGp`yQ}}G7BU{OK)GOd+@>o3ZbxH2OV8zZY{if7tLBi@ zB`}+N*ey1Qc+hAuSKN5?0u;1K_8%M%wi#LpxM_XHLI6IQrRm-d5khNLF2^(mi^u+BNa+aQ z!~1qsknCbcwRtv0H=!(sGD0IGq4QT7DW4`X!f`VDDp1% zfvH`SZyvHyf7y)zEn#QB>nddf^8M3*A6r1Cbjizh3yKB$E{l^*`QRv1ea!ASP(lUsQ6H1%j}|?M7gGPl>B=MqD!7M zNIxs_oLIAk=H?}6dBYy$B2=o3tE`+B+!KBS2%n1{pJ*m)CTB zqHqYFLN|!9IXnL(8N4z;!Xl0gB7F0R-FhL=otcS_oM$F#gN4c9NeBqq>N~16A!^M{$uP$%7`_=v-a-&{9!~ABCN53{%JK?ioc?XQLJ+&*cj~BHJ z$DyA3fyb`-RksTTu2?J<=NQbeFV;+a+{qNGF%1Tva?Uj}*suAoItkwOXk@jgv7bCB zc*)w8J#c@?Xi23=wH)=-cxS(gP7}?PF&JQD%W|DhTftw_4w*+vEaCQF=h`62sx9&l zN^)RvUd1h2!=AxMrK1(H@pp4I{k%8JCP(g9{={}67>I`6c~h~zujoZEsI>n`WN=w= z*i^M9y1OEzre7p~`P#T4MFXhgIIfS+vuoOGQCrD5W&?oeBNmjh9hp;C{Iv(s`=-E1 zSri}A$25uOix<8&$iObzE>?QoHj00{$F9N?_WE&=O_Gxc0@lK0E{4?hjWK8uQs#L7 z^j_vBdT$zIO<5V@<%J9c^uM_O>aD&fX$R%HyyApQWXDKr-8QwhvkLH&wXxUB6hubz zW#F<;MK207%hnxvF#j}_ZHk6#+EmwbE!#|#pOi3F-l!8NX7T=5?TiGT(Tp=}vPQG_ zZa424h)=pvbJuYWTL$dr~Q+wv>j1Qq*1`?snI$Z|EH1@2AFRS6wvo@+5 zFx=ItP*Wfyhey0xO-uxn!M0yq!x-&<<0-Zjt!Lq1e{~sK%NAIdwX(j}N{*yGGx*a0 z*@$W3^fL0d2D3;{wn9}*9R>J6D7{?Olt^$>F_!KYO_0bGs;MaXN|Txs*#^33>5W)3 z#8K$4q&NMdEbGa#_cM{b459MYFfD-1Nc`dK5HDVDtlK(}RFkQ3FXuVq$TUd4 zZap9-&lEf1#~zjCaW*s~VgZUPv2IfSal$f$7YJwB!8>N)5it0F2OZv#-lSZuBp=B> zQSN%?7nHJWXX~M(T;v|5&spz9aOt*B0}Qb5ufTiqH}XX1#iuT~EEckHh%4kLWbSAx z)JDUBfLu%hCns7v7PF`N2Lc45C~ER_1<#JiwfhT?x~}1FxEkLf4F#)XSgDAZOrb=t ztbQvCQ^inLyCWPU$u6k#J!(WL7;lo;+G;HWlC2z=vUuftHsbiE%b-vo3DQrvZ-HM# zJC}UY@_mf60un!~M{4>F4RPNi)n>_I71BPX`3n(4%{0p$-5?LgOy`^@)Y#Ya{F1$T z`{8~umOG$>tC7v;y0>F-u^uQl7yIWO=OKsVbW?9sS#;dpCc3XYH~bg^8!ii!CO6~9 zfT9q2sM*Bn>cFW{+{vg*7Xk$ezW&3|TiPz9@$G&YhZ6kPdMXSUjtK8G8;&yVEG-O5qB=pPaY^S|Sc6EFOp%?Yn zZ=Nn@Fu9&3Dk8}XFUt*+M>2IU`qI5WhYZsuWd z5TST?_FN zVR0**NNNFMNfRSwIio4QLO`el=I;*(sPL1_PcXJW4o$o7A4|DqxjKbKhm7pUgErkJ ziyMD3?1oRR$(SeW45DQ=Wi2votPbw*`Q!TCO?3B}>{cAmDrFS;+ksbbMxA{(oe93s zzbb=Z`|t(Wjf?#I?F0aT#{&QiY>GMU9(`iaj3v-Yt^GY8V$)_2%#Vjq6+Jm^Utk2u zx0vE!)frxik{|~%>4k72+;-FB%RI1RCvZCl78L#QNe_wVw`h^LBFbiH8s}@H|7d*n8eT7ZG;b?5pulDza=E}%MVq$ z+=)i;Iln;3w3Kx~Xf;~paHzYQnElX^v9gVBA&zYNDs*ib=JaU7gbjm|1Cyoe9gJ{wj4wqFBP z-#M5=aOGa0a4~^wrKdIsOxJQf8)1a((qS*HsLPWZEgn0vQ@!84@zhaFIX0%7h*)=e zAHk#NzmRMeKvzQ;`(c=bO6`v>7LrZg!hoN3jEI|4&uaK2g9|>oEzVhlQOn-$8({a! zLKy0n@$Tb76=o^RrzJSm-|rCKYqK$yqKQgu89ND4U9Ng3l1?JxPzS7@nl?{~)627F zhrxN&{QWg#+eT3xss1Q=S{T21Ao}e5{tRs~b!g_W3*;Puk#abk^FVQO?(}4At!>Z< zb|#z=y~h@fN!^U`g>*r%=E0pd?mtvFl@);_!$Tq;tUH8Q$5F>*ESD~C9WPyn7kco1 zxM7`&nTtZMKY+Nuka0joc7byqgj4Fyq_6nUP@C z$tXueli<-e8sff1Uf6=^4}ea#D)nyII^jdTX@zP+{ji(L`3iA`fOD-z)~)4@7)}#P z2P$cMe^Sx3iNm|zUeCr|>&^P05k-KA<;hQO<4(QdyOAUlH>-mV_MU&Y&yLfdt($h} zE~~=6K`KBo$cT;VUX#YueS@a}&SxtVQvwY_ueRb+6YI+|t_I@XbrM}49`yY6JQAVd zC5uE4X4?Zu93`5PyND0R92gTL4yTrnvL}tUe-L2ip~1&Hy2%Vc6@om(%3z^w*kgVG zxfT*NQgA@e4dPMTcg1x68rDMJAa*XiL5*7`U$lmUXvU~FfgBvr2$bxIwJcY|5}5_l z!Fq=9>w0ncYg6BCr{rG3AWX`23T{bD*6VN_CJUE;GZ-rX*;;D)Qq4J?!%z)r6>*&v zzwWRMyW8i2QCA$hoX=%y$@`^z?u`jO-mc1yDu|WVMkTIuI0bLK7}_m=qFigGUk{Sf z;`XpnDz(@u;&*k_>$R`*M*GQRtoB2L`pDg!shE z#<8W{#&ijcVz}!XwdVA*%++;PyC*{4zID7PS3&WsXorEV`*UBr3UX$&ozAk6B;m@2 zfHVZ)8w83%n=^X(3@86RXpimGKW<%;P>?(j31vCHY;#l(dJ3+Zn?}ozlH8~je=?(OD0C9S2pkpV+a^ZkT4SJ6H z-L7k=U0dgFJL-bd{`{x8VBo3u(bT$VeM}@u(lGg5w-@qH*U;a)*>r8YE;Jcbz%US) zpI07`$6C&sns%SnQ2f|cM2*w1@k2sk2Re}Sxr`AAO0t=#Eu(Zq)i~Yj>ey4N9^w64 zq@6u%@RdgIf)daGrlt`gmtR(;eT<95rI}<<(s2|jgBIy1mABz|Zf8C8k!4W0&q(4l zNY=UwTHpfc98-))d}*Tj{Oy7H@7HFF6Je$oyo^f4(w^rcD}dnN?Xcf);+{hZ%ccQN3hQW-2*gtMr|FFFSZ=>KS6@rP7QA!n)1<4#P;rLterSWwPhto{UVsFLAk;%wA! zd4o|Qw0`ZfTQ7UwLKkqnZqPRoo%f~JJCGllW1;2^d;@X@$zY5S zHaBn=4oOaXF(QIT$p5obXV%0Vs-`+UPJPdKKJAalStx`Ld&eKWRI>goW|1E%n20AY z%c``i;dOJRf1^$hhyG?F3PT~m#KUWYp+%HPd1LR8;`aKa3n4{tS#XNxCyLUK*lc-= zyQKD@Xz>|JdJ#U?LJI{7K<3<^!Ck-~qiw3~!}QXRL@E{1M|567QSy*f(?NpCGsjqi zz&+A2*P}o*O1)d`QwjCbvnlo=x1*8u8x2|LRt0K3V#%~nIDO_!jLzyr4_y^7Y<-p! zuF1!$+}VW49;fZ$#f@F`7uE?0(YWx8-?Lw_JS{#k5sV1#eLqDvi5yWlkFsa03xos89k!z?1x1KR1 zZRL`T?o+C{7ub;*bxF_#Kz|pA{nV95noUrHM6ZtjvFo=6CC;tS%ubJ(6<-P@k_H39 zSjbJ-9T^CHpT0Xv20W12O67PJ~#Z~Z~Myk2T90L!T9 zKsxS3vnU?X*h%P+oZ89?86WGiIpENS-AnmeCy{w)$@(W7zG!)8u2C&%ewZ4Dox>bF z_dM9`)N1^1#n^9~JjEom=tezZH{=b%Qsbh>o2)(|5CF_d^t?QM8n@Ap63E5fj}v{( zFxH_pCq-VkW?IatzP+I`Djj{MP=-iwn87juq@2>=5N=p(U8&pCFpY8+z^k1VsgD>K zp98S$ZTt(QWOV<=Y5DJXh)BO%)=cBE9_4IJE60c*Wm9w@)(S19URHS6p)Lp5rhn-SUiXA0PRh-?Sk;qexTTmGUZ=zh&q>y4w{VEL=a zN{GoEeA@p@kg}aAj(r_Px_3DYGFb61vYuR~lyTfK1$~@riHhBs|72i+BwWFNHMPnY z11NH=u>RSVW9e4q{BsAq1L(L4kZYrBjJ@3(kfauPth{(3hV|8Hgry`-gZNp3m*(FH z=qxUqydqb55TF(i`UqUu0=(g%=46B1FAKKt(L|Ev2*WWVN!l1n8zUiHW}X*vAG@z# zfyquNQm`r~4Qs2w?R2(W>PTX#;B1U*^_As=r6KrT2PZjx1j3wQvQWNNE0wtxQ;^Kh zlFS?_*c+RD;+W;wLh2bESU{Wj)LFA(XUHNS9W(G{9;D)WTTtZ*xn8Y0LP-~Z~`}% z!>~~5T_0ZgCRnS+sV&l&T-|*gx2yvZH;js8?FgG$y3A5XfkX(sxFZz9Z5hRYj}8A) zV@g)WvGxK)nY=ipQbwA0D0iHbD=;SoNfcEfU(iP~r! zQQzn4FMKVTc;=YYKocR+2}(wdw6f=%D{IgPa79dS^_>tGk$D}C@%_X53ty4cv+w%{O{H-p{H zuQ2`T148@{0l2>T-gz|>NeZrls*{`$6(VArfKhdG4xA9*M+P&`BtaEpo$uq0&8)m1 zp(N;6vi7^kL5vfNNihEXQ3k&1I@t&L-upMvqB2yfN$PLY_Mad$62TCfG^9BjJ&35B zRY@^=OaMSqlKS6t)&KzfcPSL$69D>e%4(p08Ug*s|5E_;KYZ|aF)$Fw zzbTG=zj=^v{-3u0m_ot-Hy-VO_ z^<57DMFK-6ViE*Lk<)`9wnJt1i_3*15vuJ)lb^jN{h@E~4+V{mfr*7pMovLVMa{y> z#?HaXB`hK;CN3cP#+`-Yw*~Qh(Js>bBI3zSIJU$^YDLExI z?N45QL19sGNoiSKeM4hYb4zPmU;n`1(D2CU*xdZW;?nZU-_@Pnz5Rp3qvMm)o7=nl zhsUSqm)C!A{e$y=(SOMPZ@7@YaRGyZf`CH&g9`}Q zs6uhMwY^XzKjg2`^zCP%(Mef$$Zr0D_Aj#kcfkDreEG(>P;xyxwN9SkcK|p~IfrT95 z8wPK4dCv5*0~X}Vv^gLxr$?;el&{?4AISN1nly8Fc}yQgbWw?3l1!vEdP)Ml*emif z55g@C8aZ_S^Bu>>|6JAY(kzuPxHvB}$R3tJSSAtg13~b;Osd;T?{RW?Vy)S9+oZ{D zUr8)Ip+S_2vb>%GVrDkr*h1L?a!+St{UxfSm>RHcQe~zgm8e16JD*4|croOYq*9Zu z?^WCUS2+D`Y8q@l^d0!2)OX zkRBm7A7W-{09!7Z_ky_!o=+BMeyf@laek$vOy`n@5n!0%zJ2p^<+p|IW9_A}?khH< z7o1Le(;IDgdxW*@dL+{8pK3Wuj&|mSs;bcMI?R8tCHe1v-y;cx6=p^K54%EkEvn&p zSDKrGZ5lvFQjv42Z0#~oMvqD^wz!!*e=g*M;&4133PFz(>UAJJ)n`NvXBX#GSp5*^ zs%bY$7BvMmBo@)EF^2YtG~j&tu&{fF0i~|ys~W{J2K`9I%Q1`>`|QE};F$dRvt)6~ zWdNVJ{RVewZbKN=7&1_bSQ8k+eYp3Yl^0DJZRCGFVFcJnKE$Ln!X|2nt-*4os?0;g z16pmtv%QyQ-`RJzZ~F;-PFpQoX6em$`I;$@|K=dK<@%+${ikuO4CC#oEsmLec>~HF z6Nob}i;N0x~~-+A-B}Z?@h{XGgDsh zv(vjlm7bN}uz5v4ZZ=#u(b7)NEnu^eg?wK5AzjflkidjnB>c$1P!}#|`(gjS!dwL} z%(M4OJeBv9nzBVTWdi?!s?~HQ&{3?V1GC*EA=A!Wl$3(u_7}p zDW2w|^b63nl?i+)z)0%pWg)U=W+`d5fcypMj|HCBhWl|INE5>$Cb56Ja`n8lIp#Cb zb;sxWNEtCI(38~jI)^igpoaU&%46HJ>LzZQY`#a0P;F#n+#Z zJye|R+&R`^V}Tek3L%N>awgCbi!2fi^;AivH_w+4nVDmV_~cAsSnDMPtcClI1X# zQ4I;HWCv`_a{k-n$dYle?I3S1@tP7qx3oWB`vc@(0Q9@7qCYf?FFa<1vNr2~9l zCxGTkKMQ9gA=9`jTEdCUx`!r{8UpV7=eO-d1#VfF7bnnPcg4HjN-tdt%(WAh0Ca{(kVM1mN~&D~J&qbZrGgF;vsM9iWV zP0dsy+4$=76^ZSa!XT@luT-rQ05VS59yet0-(NaG&SCtv&L+yMn5#iD z4IY7-@{AB1y_4j^$lZ+QJ8#T|A~PLs$UH`|AR`H|snX6G!~q6zP}E!p%v?HiCQT}* zEF3co8DwSO0>^!$`%!AB2^a>)t-GN*in(9`O%fuMNs3{@V=L0?_$F!M3)w}hylWtc zDXVL);Z+=Bxpn!=gY=-)TmksrIinBaFdaR|NcBpSo~45k48>|{?om*P86emi?`|%$ znf%`QqCVO`7dicGBgCvpcwp~iiu>t5N=&F5h_e3pm>a$^G1&H3A$md>IE;v!?pY91 zdfBM^)UrrROq$wCfc&cXlWnIU<ࢫ_vHNk$t zfY4OapNT!uclp+WR1-A^HRDpYmdIXvOCTND%)7d`4kTW~30 zKS$Ya*1_BFsgsuK;2LC$l9{{v`hB|76}e$1K=RZm#`GO3rBpUaImya+!S*S!ueVZ< zS8Zj9Mz@fEyyoZ-Dj>=+c&u`F<;Y+^|NKE*r$nt;UZ!I@2SHrxZiKbeL~VaugshU{ z!XudvFddDW`c^q8ZGKikYaUKVDyZE^&0I6;*}+M-2UJ<(jw%=PCAgX8C1QC6lah3J zm0e4}ry!1Fk>{E(_&hRF0xAA#p+R@F^YU)t5a)w(pUOS)+Ge~*n-Qw-9VXkycN)|N ze>tj-Xt~s|a-prUiG#hU)Kw`iU`B|PnYm_h`d=Y+CT`o8F7ji&%jVaO)YBOa+5g!APa{xg$AuOGa5qTaV|OF+A4a|Zj+tB29%;z;?A}u*2vfJ>igz3(2R*pQM%Uk zqKnVhqhk-O%D{Oa?B!8JgPbr1;4cpOcWGs+9dBZuSd=;5N&C>TMk;%gzqsk6+MO+5 z0k}S7+RZwV11(P8Qy9)sn+WRSjCLR2$3pZ8ws!xq5-*(IZv>- z(H*%?9C6D64?F9Ex%yT>(XC5aC59EF#lD)vN-!I!O5w9LymIDKQ)9;R&C8-j^qGPl z2xZGVKfQz@SoqJ^%4*_#8;iPWz6mVJHX-p_>Z{JgkJnXe z$}K73cGtB>D&i%tv5NSi2$bwf1CA81{Ih+TRQqN@7Fdf9?V0BMA@%3Tiq8JbcG+0+ zon--ed|;uvT4b3YjqRea15`@bDz10U`his=oD2?C=MLZYxibpWIp05b0+%9f8}gU_+q4^$8))@>`c$YcIZJW&u6qn=Oa(G6U0!gC&sLZ7Z`H=P zlwn%=nJ;M5lpbA`u@iuEyRGYi|9wePw<5lgSHfN}ai)nV4xEL4lss*lR#clq#Ha*7Dt=%XNlaCjjoZm7B4|m;c;#Fn!y+` z<^V-nauEOkI{0df2OFmpIygvd9T+Nx0#u5`CUt3h`UW4PHpA2uw5TY~W=0dgzE4d!^O7%{40?Im;ft zTV*DHcbF>=BUU$9#9pGY-Ai$Wz;220J-+eO%6a+0*>y{)LKp`kkF<=5DVTgclB50J zUcypwQA+w%6pF!_tZBSEsBU-ypzk4x^b4UEy#U`)twhN4a6MTrYGxR6Mox-)q4#Q> z*J9FQUtO7+Ze?35B5&`P+1~Pe!c4SM@>{*iU5-?7eV+Rh`)On) z+rZY=9ywKF%2I%;qJNoa%a$xG-&JL4&MYuo)*4a%dhgUrzW6)PA?1!;{Ob5GFx$Lv zaYd?duO5S#UXoE5R#22Hs6>;r;_Z{MZCr+>V=pf+?(A-Z@2Es!KaDxqhSqC^aVer; zhT>-G7r0!rbarcwzvBnb#c`fXtXm9+6r<6rxBXzPYE-Qiy0i=VkK=)Uo3{ibSb2fT zx^LoDQGkX`&Bx3y9WNpY1{8;gw?Ql8s~BrTH=ky>UScse5(4sv-TQRAeBnm*KdAWm z+~v*b?Xz?1<>;heJ24J!{uNt2@gVg9xAUw=o_SzYbwoDOIj5cphN3$DYI=SUiB)(a zT>{#=@`~#zPzs{`7esvaOtpo9Rui`0HI8qCNvo+Qv!Sn>Ak+pJ`eX&Eay$%N2zUeR zT4naH@A)m~s9+wYTi`R+zAA7JlS_syg1(}D1zsI?jtRl{ChA}(fG3$`B&S;r4A7%< zhJG{qLt{ZSAQ>(iKJ!aU-wIkQ2X#t z#xjk<*LMuZok1Xf?RE>BwbA2{4(RGsKW!o#9Oy)G6u4q4lKZ?Jbl|np5D4v&Ijmha zeBI~(@ z##SAmajLTOdCzHVU^u<^ITbshtM<|hwME#*`~PS$a~l}N`g&2aRg!of)y^i;H#a|v zI1l9l!)X&GCc_}xzm*%^v4DScnr-(8XO>%&4`e56+M> z*s`nVfGjd*Am+V+<4S&IxGc;u5C(}V|bcKp3k29 z=HCj_kPqwGN?QvN;^=%3BXmlqM$EIyOD0vQ1P@i-ZChjWKXx **YS:** Hey, have you heard of Reef Life Survey? + +Unsurprisingly, the three bots gave near-identical answers. Here are the first two sentences from each response: + +> **ChatGPT:** Yes, I'm familiar with the Reef Life Survey (RLS). It's a global marine biodiversity monitoring program that enlists the help of trained volunteer divers to collect data on reef species and habitats. [...] + +> **Claude:** Yes, I have heard of Reef Life Survey. It's an international program that involves trained volunteer divers collecting data on marine life and habitats in coastal areas around the world. [...] + +> **Gemini:** Yes, I have heard of Reef Life Survey (RLS). It is a world-renowned marine monitoring program that uses trained recreational SCUBA divers to collect data on reef fish, invertebrates, and habitats. [...] + +They all kept going with extra details, and ended their response with a suggestion to provide more information. + +## The product idea + +Once the scene was set, I hit the bots with the big idea. I sent the three the following message: + +> **YS:** I'm deeply familiar with the organisation, as I'm an RLS volunteer. +> +> What I'm after today is this: I have an idea for a web app that'd help organisations like RLS find grant opportunities and apply for them. As it relies on recent improvements in LLMs / AI (specifically text summarisation and generation), I call it MagicGrantMaker (MGM). +> +> The main value proposition for RLS is that it'd: +> +> 1. Surface new grant opportunities +> 2. Draft grant applications +> +> That is, it'd save them time and help them increase their funding. +> +> One key strength of MGM is in its focus on a specific niche for the MVP (Australian marine ecology researchers). Beyond the MVP, the idea is to keep finding under-served niches and deliver an exceptional experience to *specific* grant seekers. +> +> The second niche would be Australian renewable energy startups that may be able to get funding from the likes of ARENA. +> +> The third niche may be startups like MGM itself: supporters of climate tech and nature conservation. +> +> If this proves to be a viable product beyond a prototype, it'd follow a freemium model: +> +> * Free: Search for grants using keywords and a natural language interface +> * Paid tier 1: Create an organisation profile (much of it can be scaffolded automatically from websites and LinkedIn), and get personalised grant opportunities emailed to you. +> * Paid tier 2: Draft grant applications based on your org profile with AI. +> +> Viability would depend on data quality, coupled with the quality of the AI implementation. Then it's all up to effective marketing. The idea is that if the search works well (again, depends on getting a unique dataset), then the free part would be valuable enough to lure people in. +> +> I haven't spoken to any potential customers yet. The next step is finding grant sources to assess the viability of the search engine for the initial niche. +> +> What do you think? Please be brutally honest. + +Despite my request for the bots to be brutally honest, they still felt too gentle. What I found most interesting is that they all followed the same response structure: intro, strengths, weaknesses, next steps, and conclusion. Given the length of my prompt, I expected more divergence. + +For brevity, I'll summarise their key points rather than provide the full transcripts. + +**Strengths:** +- **[All three]** Niche focus +- **[All three]** Freemium model +- **[ChatGPT & Claude]** Scalability +- **[ChatGPT]** Value proposition +- **[Claude]** Time-saving potential +- **[Gemini]** Addresses a real pain point +- **[Gemini]** Leverages AI effectively + +**Challenges:** +- **[All three]** Data quality and access +- **[All three]** Competition +- **[All three]** AI limitations +- **[ChatGPT & Gemini]** Marketing and user acquisition +- **[ChatGPT]** Customer validation +- **[Claude]** Customer adoption +- **[Claude]** Regulatory compliance +- **[Claude]** Pricing strategy +- **[Gemini]** Freemium conversion + +**Next steps:** +- **[ChatGPT]** (1) customer interviews, (2) prototype development, (3) data partnerships, (4) AI and UX testing, (5) marketing plan. +- **[Claude]** (1) market research, (2) grant source assessment, (3) MVP development, (4) legal considerations. +- **[Gemini]** (1) data validation, (2) prototype and testing, (3) competitive analysis, (4) talk to potential customers. + +Apart from Claude being a bit too concerned about legal issues, the outputs are essentially the same. I suppose it's not too surprising: The current generation of general-purpose chatbots excel at surfacing generic advice. Often, generic advice is all you need. And as with coaching, it's the coachee who needs to do most of the work. + +## Grant sources + +My main motivation for the chat was to surface data sources and assess how hard it'd be to collect the required data. I proceeded to ask about: +- Grant sources for RLS +- Grant sources for applying AI to RLS work +- Grant sources for Australian renewable energy startups + +As with the beginning of the chats, all three recommended pretty much the same funding bodies. No surprises there! + +One thing I found interesting was the number of potential grant sources, which was in the dozens. These were all valid – I was either already familiar with them, or verified that they had funded RLS's work before. + +The number and diversity of grant sources has supported my suspicion that keeping on top of grants is a non-trivial problem. In fact, I found a potential grant to apply for as part of this work. However, this isn't enough evidence that MagicGrantMaker is a viable business idea – I'd need to talk with potential customers if I were to pursue it. Still, given the existence of similar companies in the space (e.g., a couple of startups are helping American companies get climate tech grants), I do believe that with the right niches and execution, MagicGrantMaker could be a thing. + +## Conclusion + +Paraphrasing [the words of an OpenAI engineer](https://nonint.com/2023/06/10/the-it-in-ai-models-is-the-dataset/): _"AI model behaviour is determined by your dataset, nothing else"_. Therefore, it is somewhat unsurprising to see ChatGPT, Claude, and Gemini converging to near-identical outputs. The main dataset is the open web, but a lot of additional proprietary data comes from [fine-tuning the base models on human feedback](https://huyenchip.com/2023/05/02/rlhf.html) (the part that turns auto-complete language models into friendly chatbots). + +You'd expect the proprietary data to make a difference, but companies like Anthropic and OpenAI source at least some of their proprietary data from the same providers (like [Scale](https://scale.com/)). The fact that all models compete on beating the same benchmarks also contributes to their sameness. + +While my comparative analysis is anecdotal, it has helped reaffirm my suspicion that when it comes to the top general-purpose models, it's fine to just choose one with a convenient UI and tooling. I believe that we'll keep seeing advancements, but that the top players will keep catching up with each other. This is similar to what happens in [Kaggle competitions](https://yanirseroussi.com/tags/kaggle/) and other sports, where simply knowing that a score is attainable causes all competitors to put in extra effort to attain it. + +It's a different story when it comes to special-purpose AI tools that are trained on proprietary data. MagicGrantMaker is an example of such a tool: its value would come from being hyper-niche and collecting a unique dataset from specific customers. It's simple, but not easy. And the hardest part is execution over time rather than coming up with new ideas.